This PDF file contains the front matter associated with SPIE Proceedings Volume 6833, including the Title Page, Copyright information, Table of Contents, Introduction, and the Conference Committee listing.

The users of digital cameras often take multiple photographs of the same scene. Such multiple shots usually have a special meaning to the photographer, and require further actions, e.g. selection of the best exposure/composition/portrait or stitching several images into a panorama or composite image. We present a method of fast retrieval of all groups of shots taken from the same viewpoint. This task is different from the recently emerged near-duplicate detection problem because, in our case, the multiple shots differ not only by photometric and simple geometric transformations; they can have a little or no overlap, and large variations of objects may be presented. Therefore, we solve a general multiple image registration problem by extracting local image descriptors, their matching, and recovering geometric transformation between images. Initially, the photo-collection is divided in time-based clusters, which are then refined by extracting connected components from the global image registration graph. The method has been applied to real consumer photo-collections, and we show that depending on individual camera usage styles, user collections contain from 15% to 90% of photos requiring further attention. The presented system automates the otherwise manual work of selecting a series of similar images.

A high speed real-time wavefront processing system for a solid-state laser beam cleanup system has been built. This system consists of a core2 Industrial PC (IPC) using Linux and real-time Linux (RT-Linux) operation system (OS), a PCI image grabber, a D/A card. More often than not, the phase aberrations of the output beam from solid-state lasers vary fast with intracavity thermal effects and environmental influence. To compensate the phase aberrations of solid-state lasers successfully, a high speed real-time wavefront processing system is presented. Compared to former systems, this system can improve the speed efficiently. In the new system, the acquisition of image data, the output of control voltage data and the implementation of reconstructor control algorithm are treated as real-time tasks in kernel-space, the display of wavefront information and man-machine conversation are treated as non real-time tasks in user-space. The parallel processing of real-time tasks in Symmetric Multi Processors (SMP) mode is the main strategy of improving the speed. In this paper, the performance and efficiency of this wavefront processing system are analyzed. The opened-loop experimental results show that the sampling frequency of this system is up to 3300Hz, and this system can well deal with phase aberrations from solid-state lasers.

The two dimension performance model for low light level (LLL ) imaging systems will be studied and built in this paper, which is based and improved on the Johnson criteria and can also be used for thermal imaging systems. Two dimension threshold resolution angle and universal apparent distance detecting equation for LLL imaging systems will be given out. The new number of the resolvable circles across the target and the background for detection, recognition and identification will be put forward and some different definition for Johnson criteria will be discussed. Two dimension performance model is more accuracy and nearer practice than one dimension models especially for the case of the length is much larger than the width of the target and the background.

This paper concentrates on images formation simulation under low light level condition (10^-6-3lx) and photon limited condition (<10^-6lx). In the first part, we introduce the main characteristics and features of low light level images and system entire noise and simulate a deblurred image intensified by photon imaging system recently constructed under low light level condition. The influence of scene luminance and photon imaging system optical errors on the simulation is introduced. Then the system entire noise is appended to low light level images by a novel noise analysis and generation method based on experimental study method. The second part of this paper deals with simulation of photon images. Because of randomicity of photon images, roulette wheel selection is utilized to confirm the grey level of stochastic signal photon image and noise photon image is generated by poissson stochastic process pixel by pixel. The final photon image is acquired by synthesizing the two images. The simulation presented in this paper provides an economical and convenient method to investigate the detection ability of photon imaging system and image reconstruction algorithm under low light level condition and photon limited condition.

An automatic optical inspection algorithm for chip components which is based on local principal wave probability is presented. The gray-level change of component image in local feature region is used as a detection criterion. By introducing local principal wave probability, the similarity between sample image and inspected image is described by intrinsic characteristic of the principal wave. Thus it can be decided whether the component is a defective one. Furthermore the type of defect can also be concluded. Experimental results show that the algorithm can be real-time by reducing the computation from the whole image to local region. It solves the problem of high sensitivity to position of component that exists in the algorithms widely used in automatic optical inspection for chip components.

Video systems have been widely used in many fields such as conferences, public security, military affairs and medical treatment. With the rapid development of FPGA, SOPC has been paid great attentions in the area of image and video processing in recent years. A network video transmission system based on SOPC is proposed in this paper for the purpose of video acquisition, video encoding and network transmission. The hardware platform utilized to design the system is an SOPC board of model Altera's DE2, which includes an FPGA chip of model EP2C35F672C6, an Ethernet controller and a video I/O interface. An IP core, known as Nios II embedded processor, is used as the CPU of the system. In addition, a hardware module for format conversion of video data, and another module to realize Motion-JPEG have been designed with Verilog HDL. These two modules are attached to the Nios II processor as peripheral equipments through the Avalon bus. Simulation results show that these two modules work as expected. Uclinux including TCP/IP protocol as well as the driver of Ethernet controller is chosen as the embedded operating system and an application program scheme is proposed.

Shape Distribution is fast, simple, and robust method in 3D model retrieve. This method, however, only considers distances between the objects' shape distribution histograms and ignores the information included. As the result, the retrieval precision is low. To enhance the retrieve efficiency, a novel method which integrates Shape Distribution and Self-Organizing Feature Map (SOFM) is proposed. The models' shape distribution histograms are established by Shape Distribution and transformed into the proper format of SOFM. The similar models are grouped in neighboring neurons of SOFM by using competitive learning approach. In addition, the dissimilar models are indexed in far away neurons. With the given query model, SOFM classifies it into the proper cluster and exports the retrieval results. A case study is presented and the results show that the retrieval precision of the proposed method is higher than that of the Shape Distribution method.

Medical ultrasonic Doppler imaging is one of the most important domains of modern medical imaging technology. The application of coded excitation technology in medical ultrasonic Doppler imaging system has the potential of higher SNR and deeper penetration depth than conventional pulse-echo imaging system, it also improves the image quality, and enhances the sensitivity of feeble signal, furthermore, proper coded excitation is beneficial to received spectrum of Doppler signal. Firstly, this paper analyzes the application of coded excitation technology in medical ultrasonic Doppler imaging system abstractly, showing the advantage and bright future of coded excitation technology, then introduces the principle and the theory of coded excitation. Secondly, we compare some coded serials (including Chirp and fake Chirp signal, Barker codes, Golay's complementary serial, M-sequence, etc). Considering Mainlobe Width, Range Sidelobe Level, Signal-to-Noise Ratio and sensitivity of Doppler signal, we choose Barker codes as coded serial. At last, we design the coded excitation circuit. The result in B-mode imaging and Doppler flow measurement coincided with our expectation, which incarnated the advantage of application of coded excitation technology in Digital Medical Ultrasonic Doppler Endoscope Imaging System.

We present a scheme of integrated optical 3-D digital imaging (IO3DI) based on digital signal processor (DSP), which can acquire range images independently without PC support. This scheme is based on a parallel hardware structure with aid of DSP and field programmable gate array (FPGA) to realize 3-D imaging. In this integrated scheme of 3-D imaging, the phase measurement profilometry is adopted. To realize the pipeline processing of the fringe projection, image acquisition and fringe pattern analysis, we present a multi-threads application program that is developed under the environment of DSP/BIOS RTOS (real-time operating system). Since RTOS provides a preemptive kernel and powerful configuration tool, with which we are able to achieve a real-time scheduling and synchronization. To accelerate automatic fringe analysis and phase unwrapping, we make use of the technique of software optimization. The proposed scheme can reach a performance of 39.5 f/s (frames per second), so it may well fit into real-time fringe-pattern analysis and can implement fast 3-D imaging. Experiment results are also presented to show the validity of proposed scheme.

With the rapid development of the electronic technology, multimedia technology and network technology, video monitoring system is going to the embedded, digital direction. In this paper, a solution of embedded video monitoring system based on OMAP5912 is proposed. This solution makes full use of the advantages of dual-core of OMAP, which include ARM core and DSP core. Non-realtime task such as user interface and control unit is assigned to ARM, and realtime task such as video encoding is assigned to DSP. The capture and control task of the ARM side and the video encoding task of the DSP side are described in detail. The experiments demonstrate that the video encoding speed has been greatly improved by the proposed system comparing with the single ARM chip system. The frame rate of the monitoring system is increased in a large scale, and more suitable for the application in practice.

We use the conception of pixel-matched spread function (PMSF) to analyze the physical process of intra-page cross talk in a holographically imaged data page, and design an arithmetic to suppress the intra-page cross talk generated in the reconstructed data page due to limited aperture of imaging optical system. By applying this arithmetic to a data page of 512×512 pixels captured from our CCD, the raw bit error rate (BER) decreased from 1.1×10^-3 to 2.4×10^-4.

Theoretically speaking, image quality assessment algorithms are design to evaluate all kind of distorted images. However, as the JPEG compressed image dramatically distinguished from images with white noise, every distortion type has its own characteristic. A new method is proposed based on the characteristic of JPEG compressed image, witch correlated well with the subjective result.

An integration system was designed to apply to the remote communication of optics and electronics detection systems, which was integrated with programmable DSP and FPGA chirps in addition to a few Application Specific Integrated Circuits (ASICs). It could achieve image binarization, image enhancement, data encryption, image compression encoding, channel encoding, data interleaving, etc., and the algorithms of these functions might be renewed or updated easily. The CCD color camera being a signal source, experiments had been done on the platform with a DSP chirp and a FPGA one. The FPGA chirp mainly realized the reconstruction of image's brightness signal and the production of various timing signals, and the DSP chirp mainly accomplished the other functions. The algorithms to compress image data were based on discrete cosine transformation (DCT) and discrete wavelet transformation (DWT), respectively. The experiment results showed that the developed platform was characterized by flexibility, programmability and reconfigurability. The integration system is well suitable for the remote communication of optics and electronics detection systems.

A new image corner detection algorithm is proposed in this paper based on the SUSAN algorithm for the electronic image stabilization of the UAV video image. Through analyzing the gray characteristics of the image of the UAV, the new algorithm changed the judge criteria of the SUSAN corner detection algorithm to increase the accuracy and velocity of the image processing. The basic steps of the algorithmic show as follow: First, the correct threshold is decided using the gray characteristic. Second, comparing the one pixel with the eight neighborhood pixels, the elementary direction of corner is acquired. Last, a corner is acquired through calculating the number of the congener pixels based on these new directions. Using this new algorithm the corners should be detected fast and efficiently. Experimental results of the UAV video image processing show that the new method can highly increase computational velocity. Consequently the proposed algorithm meets the need for real-time image processing.

Due to the affection of large moving object in the video image, the ordinary image stabilization algorithms can't get precise motion vector of the image. In this paper, a new image stabilization method that explicitly deals with video images containing large moving object is given out. It detects a rough area containing moving object first. Then this area will be got rid of from source image. Finally, a feature area taken from the rest part is used to calculate the image motion. For the affection of moving object has been eliminated, motion vector's precision of the image is improved greatly.

Speckle noise reduction is an important technique to enhance the quality of ultrasonic image. In this paper, a despeckling algorithm based on an adaptive block-based singular value decomposition filtering (BSVD) applied on ultrasonic images is presented. Instead of applying BSVD directly to ultrasonic image, we propose to apply BSVD on the noisy edge image version obtained from the difference between the logarithmic transformations of the original image and blur image version of its. The recovered image is performed by combining the speckle noise-free edge image with blur image version of its. Finally, exponential transformation is applied in order to get the reconstructed image. To evaluate our algorithm compared with well-know algorithms such as Lee filter, Kuan filter, Homomorphic Wiener filter, median filter and wavelet soft thresholding, four image quality measurements, which are Mean Square Error (MSE), Signal to MSE (S/MSE), Edge preservation (β), and Correlation measurement (ρ), are used. From the results, it clearly shows that the proposed algorithm outperforms other methods in terms of quantitative and subjective assessments.

The tracking and orientation of optoelectronic targets must obtain the data of target's velocity and angle by prediction algorithm. But the state and measurement equations are usually nonlinear and uncoupled models, so the tracking problem often connects with nonlinear estimation. The commonly classical extended Kalman filter (EKF) algorithm suffers from a lot of defects. There are those problems such as easy to diverge and the convergence rate is slow and the tracking accuracy is low. In this paper, a new nonlinear adaptive Kalman filter (AEKF) algorithm based on the adaptive tracking theory in current statistical model is presented. It expresses variation of acceleration with the information of position and angle to carry out self adaptation of noise variance in on-line mode, and to compensate the linear errors of model in dynamic mode. Analytic results of Monte Carlo simulation prove the AEKF algorithm is right and feasible, and the accuracy and the convergence rate are both improved. It has better performance than the EKF algorithm and modified variance EKF (MVEKF) algorithm in the tracking and orientation of optoelectronic maneuvering target. The simulation results and new method will been widely and directly applied into various engineering.

It is a trend in the world now that uses state maintain method to maintain the jet plane's engine by taking advantage of the advanced UV sensors and UV CCD camera. Enrichment phenomenon is one of the most popular plane engine failures, and it has large damage to the plane engines. In this paper, a new technology will be introduced to determine plane engine's enrichment phenomenon by using UV image processing method. The new technology acquires the enrichment plane engine's UV images and then extracts the features by analyzing those UV image's histogram and other histogram's characteristics just like the mean value, the maximum value, the minimum value and the standard deviation of grey scale.

The proper segmentation of the iris image determines the iris recognition accurate to a great extent. Most of the iris images are covered by upper or lower eyelids, thus it is essential to detect the eyelid boundary for improving the iris recognition accuracy furthermore. An eyelid detection method based on maximal connection path is presented in this paper. After the preprocessing of the iris image, the horizontal segmentation operator and image binarization are used to extract the eyelid edge information. The eyelids span the whole image in horizontal direction and the average of vertical gradients is larger in the area with eyelid boundary, therefore, the horizontal distance of the connection area with eyelid boundary should be the longest one in the edge image. In use of this feature, the candidate edge points of eyelid boundary are detected. Eventually, the eyelid boundaries are modeled with the parabola curves. The algorithm performance is tested in CASIA Iris Database, and experiment results show that about 0.117 second at speed and 88.9% at precision are reached for the upper eyelid detection, and about 0.078 second at speed and 98.5% at precision for the lower eyelid detection. In comparison with Daugman's method, this algorithm enhances the detection speed largely and shows good accuracy.

A method of measuring the key sizes of a human 3D mesh model was presented. In the proposed method, human model postures were adjusted according to the measuring reference plane, measuring direction and measuring points. Then the skeleton of the model was extracted through the gradient rapid descent algorithm of the radial base function. What follows was to locate the joints and set their freedom and active ranges. Then, the surface peaks were bonded to the joints through flexible models to implement human animation. In this way, the standing and sitting poses were obtained. After automatic means of body characteristic points location had been studied, the points and plane were extracted to locate the measuring points precisely. Experiments were carried out on a man respectively and their key sizes were shown.

In airborne synthetic aperture radar (SAR), the arbitrary aircraft maneuvers significantly degrade the achievable performance of the imaging system. Compared with other SAR processing algorithms, the time-domain subaperture processing is tested to be more effective in handling the problems, where the real-time adjustment of the radar parameters is required. In this paper, a subaperture image formation for high-quality SAR imaging without changing any parameters is presented. Some new characters are briefly described. Sets of the simulation results are then given as demonstrations to show its effectivity.

A new sub-pixel interpolation algorithm based on multi-resolution filtering technology is presented in the paper. During the course of the analyzing the algorithm, we theoretically testified the conclusion that the sub pixel imaging technology can improve system image resolution. It can be concluded from the experiment result that the new interpolation algorithm can raise the resolution remarkably. We presented a real-time system based on TI DSP to realize the new algorithm. The sub pixel imaging system is developed with the flexible whole design and modular design so that it has the characteristics of extending, maintainability and easy soft developing.

Digital image inpainting is an interesting new research topic in multimedia computing and image processing since 2000. This talk covers the most recent contributions in digital image inpainting and image completion, as well as concepts in video inpainting. Image inpainting refers to reconstructing the corrupt regions where the data are all destroyed. A primary class of the technique is to build up a Partial Differential Equation (PDE), consider it as a boundary problem, and solve it by some iterative method. The most representative and creative one of the inpainting algorithms is Bertalmio-Sapiro-Caselles-Bellester (BSCB) model. After summarizes the development of image inpainting technique, this paper points the research at the improvement on BSCB model, and proposes two algorithms to solve the two drawbacks of this model. The first is selective adaptive interpolation which develops the traditional adaptive interpolation algorithm by introducing a priority value. Besides much faster than BSCB model, it can improve the inpainting effects. The second takes selective adaptive interpolation as a preprocessing step, reduces the operation time and improves the inpainting quality further.

The mean shift algorithm is an efficient technique for tracking 2D blobs through an image. The scale of the mean shift kernel is a crucial parameter. Classic Mean shift based tracking algorithm uses fixed kernel-bandwidth, which limits the performance when the object scale exceeds the size of the tracking window. Although some modified algorithms can settle the problem of object zooming in a way, these algorithms are helpless to the object rotation. Based on the analysis of the scale-space theory and the current Mean shift algorithms, a scale and rotation adaptive mean shift tracking algorithm is proposed. Experimental results show that the new method can effectively and accurately obtain the best description of the target areas for the first frame, and the new mean shift tracking algorithm can adapt to any kind of object's movements.

A new detection method for unstructured road based on robot's vision is proposed to improve the effectiveness of road detection in complex environment. In this article, the OTSU, an auto-adapted threshold searching algorithm, is mainly used to classify the road images. Meanwhile, to solve the problems of misclassification in complex environment, the OTSU will be used the second time to subdivide. And multiple scene templates are built combining road referring window (RRW). Then, multi-dimensional features are chosen for region reorganizing according to those templates to obtain the optimal classification. At last, the classifying results are merged by referring RRW to extract the final road region accurately. This algorithm shows good self-adaptive ability and only needs little priori knowledge. It is also robust against noises, shadows and illumination variations and shows good real-time performance. It has been tested on real robot and performed well in real road environment.

In this paper, by analyzing the basic road features in remote sensing images, the model of road extraction is discussed. The popular methods of road extraction and their advantages and disadvantages are generalized. The recent progress and results of our research group at relative aspects are introduced. The development of the issue is also presented.

In this paper, a new method to recognize bridge in the complicated background is presented. The algorithm takes full advantages of the characteristics of the bridge image. Firstly, the image is preprocessed and the object edges are extracted. Then according to the limitations of traditional Hough transform (HT), the extraction method of the image line segment characteristic of HT is improved, which eliminates spurious peaks on the basis of global and local thresholds, discriminates the position relation between two straight line segments, and merges segments with near endpoints, etc. Experiments show that this algorithm is more precise and efficient than traditional HT, moreover it can provide a complete description of the bridge in a complicated background.

A novel method for the coarse registration of range images is proposed. This approach is based on texture-feature recognition. As the development of optical digitizing technique, it is now able to acquire the range images and associated texture images sequentially or simultaneously. It's possible to identify the range feature points through texture feature points. Scale Invariant Feature Transform (SIFT) is an efficient method for texture feature generation. SIFT transforms texture image into a large collection of local feature vectors, each of which is invariant to image scaling, translation, and rotation. The mismatched correspondence pairs can be discarded using random sample consensus algorithm based on epipolar geometry constraint. We select more than three well-registered texture-feature pairs, with which we could find the associated range-feature pairs of the range images. Initial pose estimation of the two involved range images can be computed by these range pairs, and the fine registration is implemented using iterative closest point (ICP) algorithm. Our approach utilizes the texture information to register the range images, leading to a technique that can be automatically performed while the influence of 3D noise can be avoided. The experiment results demonstrate that the proposed approach is efficient and robust for the registration of multiple range images.

Nowadays seals play an important role in China. With the development of social economy, the traditional method of manual check seal identification can't meet the need s of banking transactions badly. This paper focus on pre-processing and registration algorithm for check seal verification using theory of image processing and pattern recognition. First of all, analyze the complex characteristics of check seals. To eliminate the difference of producing conditions and the disturbance caused by background and writing in check image, many methods are used in the pre-processing of check seal verification, such as color components transformation, linearity transform to gray-scale image, medium value filter, Otsu, close calculations and labeling algorithm of mathematical morphology. After the processes above, the good binary seal image can be obtained. On the basis of traditional registration algorithm, a double-level registration method including rough and precise registration method is proposed. The deflection angle of precise registration method can be precise to 0.1°. This paper introduces the concepts of difference inside and difference outside and use the percent of difference inside and difference outside to judge whether the seal is real or fake. The experimental results of a mass of check seals are satisfied. It shows that the methods and algorithmic presented have good robustness to noise sealing conditions and satisfactory tolerance of difference within class.

An algorithm for image matching is proposed, which uses both epipolar and homography constraints. At the first step, the Forstner algorithm is employed for features extraction. The features description and matching method of SIFT are used to find a group of original correspondences, then the correspondences are refined by LSM(Least Square Matching). With the refined correspondences the RANSAC algorithm estimates the fundamental matrix robustly and the more accurate correspondences with less outliers are gotten, which are called as correspondences candidates. As the features extracted from the image are all the edge inflexions, texture nodes with maximal intensity or corners of the objects in the 3D world. The features which are adjacent can form a local plane or quasi-plane. So the homography constraint is proposed for image matching. At the second step, the corresponding features seeds around the feature to be matched are recognized from the correspondences, which associate with a real 3-D scene plane or quasi-plane. Then with the seed correspondences the local homography matrix is computed. At last, under the guide of the local homography matrix, the coarse position of the target feature on the opponent image is found, then with the constraints of the epipolar line and the coarse position, the normal correlation and LSM matching methods are employed to match the features accurately. The algorithm searches for the corresponding feature only in a very small region and works quickly. Experimental results show that the algorithm is efficient and it improves the robustness and accuracy of the automatic image matching.

To relieve the inefficient of traditional method on auto car tire character recognition based on hand-copying, in this paper, we present an intellective method of fast recognition. Several algorithms were involved in this method, including polar transformation, improved canny operator and distance fast cluster. The procedure was consisted of several steps. In first step, tire image was rotated by polar transformation and mosaic to form a rectangle area by Bilinear Interpolation algorithm. Compared to other methods this part can reduce system time expense greatly. Next, Histogram Equalization algorithm was used to improve the gray distribution and image contrast, which always low on the foreground and background of image for the curing process. Then, the real edge was extracted by improved canny operator, which has advantages of smoothing image, binarizing image and suppressing the big noise caused by rough surface. Subsequently, based on foregoing results, tire image was divided into several small regions, and the judgment of each region whether belongs to real character region was processed by utilizing the continued similarity and the ratio of black pixel number to white pixel number. Furthermore distance fast cluster was utilized to filter noise and split image, and class merging was also used to complete character split. Finally, character feature vectors were extracted and character pattern recognition was completed using them. This method was tested on a series of experiments, the result shows small time expenses and high recognition ratio, which demonstrated that this method can satisfy the special requirement of fast and effective recognition of autocar tire character.

Niche Genetic Algorithm (NGA) is proposed to recognize a disconnected nonparametric curve from a noisy binary image. The fitness function used in the NGA is derived from the hypothesis: Human Visual Tradition Model (HVTM). Sharing function based niche technique and elite-preserving strategy are utilized to preserve population variety for converging at the global optimum. It has the advantage of using a nonparametric method to extract disconnected curves from the noisy binary image other than the parametric method, which Hough Transform (HT) can conclude. The curve extracted by using the nonparametric method is verified by comparing the best strings respectively along rows and columns in the permutation-based encoding space. The curve length can be derived automatically from the image by calculating the accumulation of the distance between the neighbor tiles in the extracted curve. In this paper, it is analyzed that the odd order moments of the tiles in the raw image is more sensitive to the tiles with cracks other than tiles with only noise, the algorithm complexity is sensitive to encoding approach, the evolution converge characteristics are sensitive to sharing function and parameters in fitness function. Experimental results present that the approach was successfully used in pavement crack detection.

Group actions play a key role in intelligent meetings. In this paper, we propose a probabilistic approach to online segment meetings as a sequence of group actions, such as monologue, presentation, discussion, and break. In our approach, we decompose group actions into three sub-actions according to three sorts of features independently: audio features, video features and group visual features. In accordance with this assumption state spaces are decomposed into two levels of resolution: meeting actions and meeting sub-actions. Multi-stream dynamic Bayesian network is constructed based on three level state nodes modeling group actions, sub-actions, and three sorts of features. Particle filters are applied to efficiently online recognize group actions, which is based on the estimate of joint posteriors over node states of multi-stream dynamic Bayesian network. Posterior probabilities over all state spaces are represented by temporal sets of their weighted samples. We make seven compared experiments with the different sample numbers of 50,100, 200,300,400, 500 and 600. The recognition accuracy gets higher when there are more sample numbers, but is takes more time for event inference.

Since the natural rocks have quite different textures even they are in the same class, it is very difficult and challenging task to classify each type of natural rocks. In this paper, we present a method to classify each type of rocks using the modified version of Spatial Frequency Measurement (SFM). In our approach, each type of color rock images are firstly transformed into two dimensional intensity features, obtained from the highest and lowest eigenvalues of the Principle Component Analysis (PCA). The highest and lowest eigenvalues are corresponded to the most and least significant feature components. Next, the textural contents of each component are measured using the modified version of SFM, which measures all overall activity level of each component in two directions including vertical, horizontal directions by shifting one by one pixel for two-neighborhood pixels in both direction. Before applying modified version of SFM, the edge detection operator, Sobel operator, is applied to the most significant component only. After applying the modified version of SFM to both components, two textural features are used to define each type of rock. In our experiments, we test our approach to classify on 14 different classes of rock textures, each class has 30 samples. From the results, we found that the scatter plots of each type of rock features are obviously grouped and stuck together in the same class while the different classes are clearly separated.

For the optimal weighted coefficients of per frame fused video sequences image are variable, the method of PCA is applied to automatic adaptive selecting the weighted coefficients of the low frequency layers which are decomposed by the Laplace pyramid. For the sake of enhancing the video image contrast, a partial contrast method based on the fusion formula of the high frequency layers of Laplace pyramid is presented in this paper. Finally, some measures are presented to enhance the fused video effect based on the characteristic of each layer of the Laplace pyramid. Experimental results show that our method is effective.

In this paper, a novel method is proposed to implement Laplacian pyramid image fusion on FPGA. Firstly, implementation of image fusion algorithm based on Programable DSP (PDSP) and FPGA is compared, as well as the advantages of Laplacian pyramid for parallel processing. Secondly, the architecture and characters of Laplacian pyramid is analyzed in detail. Finally the related logical modules in FPGA are designed according to their functions of this algorithm, including controlling module, decomposing module, fusion module and reconstruction module. Inside the decomposing module, 3-stage pipeline is designed for decomposing images at each level. Three-level Laplacian pyramid image fusion algorithm is adopted through Verilog Hardware Description Language according to the designed methods forementioned. The design is verified on a real-time dual-channel image fusion system based on Virtex-4 SX35 FPGA. The experiment results show that the fusion system can realize real-time image fusion processing for dual channels 640×480 images at the rate of 25 frames per second. Comparing with input digital video stream, the output video stream delays less than 10 horizontal line clocks.

An image fusion method using the quincunx sampling lifting wavelet transform combined with the fusion strategy of area edge change is put forward. Lifting wavelet transform can realize fast computation and no auxiliary memory, which could realize integral wavelet transform. Quincunx sampling adopts the scheme suitable for visual system and has the non-rectangle segmentation spectrum. Quincunx sampling lifting scheme, which is separable wavelet, combines both of their advantages. Furthermore, the fusion strategy of horizontal, vertical, diagonal edge change for low frequency image could reserve object integrality of source image. At the same time, the algorithm complexity and system Input/Output are calculated, after which the small integrated dual-spectral image fusion system with TMS320DM642 DSP as its kernel processor is then shown. As the hardware design of function, principle, structure and high speed circuit PCB design is presented, software design methods and implementation on this fusion platform are simultaneously introduced. The dual-spectral image real-time fusion system is built with high performance and small board dimensions, which lays a solid foundation for future applications.

Image segmentation is to partition an image into meaningful regions. An image segmentation approach based on chaotic ant colony algorithm is presented in this paper. The approach performs the image segmentation by selecting the optimal threshold values, where the multi-threshold values are used. First of all, an entropy function corresponding to an image is defined. The optimal threshold values are obtained by making the entropy function reach the maximal value. Secondly, an approach based on ant colony algorithm is presented for the computation of the optimal thresholds. In order to improve the computation performance of ant colony algorithms, for example, to avoid the algorithm search being trapped in local optimum, we use chaotic approach to find a better solution whenever all the ants have finished the operations. The chaotic approach searching the space around the ant which is the best so far. Besides, the initial solutions are generated by chaotic approach, this improves the quality of initial ants. The experimental results show that the approach proposed in this paper can get the near optimal threshold.

This hybrid segmentation algorithm presented is a combination of three traditional methods. It has some advantages that traditional methods don't have: first, the edge detected is continuous; second, segmentation result is accurate; third, overgrowing doesn't exist. The hybrid algorithm has been implemented on some images. Through the segmentation result, it is proved to be effective.

A color image of digital camera is obtained by interpolated CFA (Color Filter Array) data from single sensor digital camera. Using JPEG compression method to interpolated data, the maximal compression ratio is less than 90:1. In this paper, an improved compression method is discussed which applied JPEG compression to raw data after two luminance components had been processed further. Results show that CPSNR of improved method is the highest among conventional method and existing methods while the compression ratio is over 90:1. And another attractive result is the image quality is better when the compression ratio up to 120:1, and the conventional JPEG compression method is helpless at such compression ratio. The further work will be concentrate on its application with hardware support.

Realistic image rendition, concerning on color constancy and lightness, is usually qualified by subjective evaluation, involving uncertainly psychophysical course. Whether common objective image quality metrics can be adopted to evaluate the rendition results is studied in this paper. Several common objective image quality metrics such as RMES, PSNR, and a newly universal one named Q metric are introduced. Experiments show that it's applicable to take images under standard lighting conditions (e.g. D65) as reference images, which is necessary for objective assessment. Experiment results confirm that introduced objective image quality metrics can help to evaluate the lightness and color constancy ability of images, in case of taking images under standard lights as reference images.

A color transfer scheme for visible and infrared images is presented. Two main procedures are included: image fusion using steerable pyramid in YUV color space, color transfer based on local mean value of infrared image to enhance hot contrast. Firstly, visible and infrared images are decomposed into 18 subband images with a 4-scale 4-oriatation steerable pyramid that contains one highpass subands, one lowpass subband and sixteen bandpass subbands. In each suband image, Y component of the fused image is formed by the pixels whose value is the larger one between the visible and infrared images. The weighted subtracting operations between visible and infrared constitute the U and V components. Then, during the process of color transfer, the local gray mean in the 5×5 window of the infrared image is concerned. The V component that represents the difference between luminance(Y) and red color is increased by the ratio of the local mean value to the global mean value. Therefore, the hot contrast of infrared is enhanced by rendering hot targets intense red color. Test results show that, the image fusion with the 4-scale 4-oriatation steerable pyramid multiples the paths to transfer the color and luminance of a target image into the fused images, thus, the transferred images are much more colorful, and synchronously reserve the two image's advantage that the visible image is good at situation awareness and the infrared image is superior in target detection.

Multiple Region of Interest (ROI) coding is important in applications where some parts of the image are of higher importance than others and need to be encoded at higher quality than the background. The new image coding standard JPEG2000 recommended the Maxshift method to complete ROI coding. However, the drawback of the Maxshift method is that the coefficient bitplanes of all ROIs must be scaled with the same values, which cannot code ROIs according to different degrees of interest. This paper describes a flexible multiple ROI coding scheme called BTShift (Bitplane Twice Shift). The proposed method uses different bitplane scaling strategies between low frequency subbands and high frequency subbands. Experimental results on several state-of-the-art images show that the new scheme does not only support multiple ROI coding according to different degrees of interest, but also can encode the ROIs and background without their shape information. Additionally, BTShift is compliant with the recommended Maxshift method. So we hope the proposed method can be valuable for the remote sensing image compression and medical image coding in the future.

Recently, stable processes have turned out to be good models for many impulsive signals and noises. The speckle noise in underwater, SAR and the cosmic background images has been proved to have heavy tails distributions and Long Rang Dependent (LRD) structures. In this paper, the Fractional Levy Stable Motion (FLSM) is introduced to model such speckle phenomenon. The synthesis approaches employing Random Midpoint Displacement (RMD) and FFT technology are presented to generate such speckle image respectively. Then, we introduce Wavelet Analysis (WA) method to estimate the LRD exponent H and propose two new technologies in estimation H parameter by Fractional Low Order Moment (FLOM) and Fractional Spectrum (FS) respectively.

Micro-satellite is characterized by miniaturized structure and low cost, so it is a good choice to use area array CCD space camera as image system on micro-satellite. FT-18 is a monochrome frame transfer image sensor offering 1024×1024 pixels with excellent antiblooming and variable electronic shuttering. The main components of driving circuit for FT-18 include power supply unit, microcontroller unit, clock signal generator unit, and analog-to-digital (A/D) converter. The microcontroller unit controls startup sequence of all voltage, the exposure time of CCD and the working status of A/D converter; the clock signal generator unit generates sequence signals for CCD and A/D converter; the A/D converter converts the output of FT-18 to a 12-bit digital output. Special attention should be paid to the reliability of this camera for it will work in a condition different from ground. The camera may suffer from vacuum discharge, particle radiation, strong shock, hypergravity and so on. All these should be considered in the design of space camera, and enough environment tests should be done to ensure it can work normally in space.

A novel non-rigid registration algorithm within multi-resolution block matching framework is presented for accurate and robust image registration in the presence of incomplete image information. After getting the deformation field computed from block-matching, we introduce robust and structure-adaptive normalized convolution in spatial regularization of deformation field. Unlike traditional framework of normalized convolution, in which the local deformation is modified through a projection onto a subspace, however, the applicability function of structure-adaptive normalized convolution based on an anisotropic Gaussian kernel is adapted to local linear or edge structures in the images to be registered. This leads to more samples of regions of homogeneity being gathered for the regularization of deformation field, which can reduce deformation diffusion across discontinuities. A robust signal certainty is also adapted to each displacement vector in the deformation field to measure its accuracy. The results show that the method is sufficiently accurate and robust to incomplete image information for multi-temporal non-rigid image registration.

The paper presents a new Adaptive Gain calculating approach when using the Adaptive Image Enhancement algorithm based on Wavelet Transform. The basic technique is to select two different thresholds which divide the input into three parts after the wavelet coefficients are normalized. For the wavelet coefficients less than the smaller threshold, just make the output zero. And this provides a de-noise effect. The output remains unchanged when the wavelet coefficients are greater than the larger one. For the last part, find a function to make the output figure S shape. So the algorithm can give a clear contrast and the function is the key. The final goal is to use this method in the process of Online Vision Measure, we have chosen the TI TMS320 DM642 Digital Signal Processor or DSP because of its powerful multimedia processing capability. What's more, the TI corporation has provided a variety of software develop libraries as well as the 3rd party tools. All such tools make the development more rapid and convenient. After the technique this paper provides is implemented on DSP, a series of optimization will be performed to make it suitable for industry real-time usage.

A new PSF estimation algorithm based on the wavelet theory is put forward. First, restored images are obtained by using a non-blind regularized deconvolution technique for a wide range of values of the PSF parameters. Then, a wavelet-based criterion is proposed for choosing the optimal PSF parameter. Experimental results show that the images restored by the proposed algorithm are visually improved, and the estimated values of parameter are close to the actual values. The differences between them are less than 1 pixel and the BSNR has been improved about 2db.

The important function and signification of researching on performance model for photo electronic systems is discussed, and the emphasis is on the thermal imaging system models for their technical characteristics and developing history. Based on analyzing and studying existing detection theory and performance models of thermal imaging systems limitations and deficiencies are pointed out. The proposals for establishing the new threshold detection theory, apparent distance detecting equation and performance model are put forward in this paper, which will be based on the quantum noise fluctuation theory and linear filter theory, the threshold characteristics and visual theories of the human eye, and can been widely used in the first and second generation thermal imaging systems.

A touch-panel is display overlays which have the ability to display and receive information on the same screen. The advantage of this touch screen is that it is easy for all users to operate intuitively. In addition, a touch-panel interface is utilizable for multi-users. However a conventional system cannot provide us with direct touching in the air because the touching point differs from the actual displaying space. The reason is that a conventional touch-panel system detects the user's operation on the display screen. In the virtual 3D space, it is important to realize that the user can operate at the same space. The authors developed a prototype virtual air touch interface system for interaction in the virtual 3D space. In this paper, we propose the interface system using a theremin which is a musical instrument having the unusual aspect of being controlled by the performer's hand motions near the antennas.

There are various displays, for instance, a computer CRT or LCD monitor, a 3D TV monitor and so on. It is possible for the display unit to provide the same size of displaying area as the image screen on the panel. Thus the conventional display can show only one screen, but it is impossible to enlarge the size of a screen, for example twice. In this paper, we present an enlarging method of display area using a liquid crystal layer. Our extension method enables the observers to show the virtual image plane and to enlarge a screen area twice.

The authors developed a prototype information query system. It is easy to get the information about an object with in your reach. But it is troublesome to do the same in case that the object is far away. If someone is around you, you can ask an easy question with a finger pointing; "What is that?" Our developed system also realizes this approach using information technologies. The system consists of a laser pointer, transmitter and receiver units for an optical communication. The laser pointer is used for pointing an object. Moreover this laser light is modulated for sending information about user's identification (ID) codes to identify who asks a question. Each object has a receiver for laser light communication and sends user's identification to a main computer. After pointing an object, a questioner receives an answer through a wireless information network like an email on the cellular phone.

The authors developed a prototype video viewing browser. Our video viewer has a function to playback movies on the WWW according to the playing scenario. This scenario makes new scenes from original movies. Our video browser features this scene scenario where you can arrange movie's video clips, insert transition effects, apply colored backgrounds, or add captions and titles. The video movie contents on the WWW are copyrighted. The browser cannot alter web's movie contents owing to its copyright like that a conventional video editing software adds effects to the original. The editing software produces reproductions, but our browser doesn't. The browser adds effects according to the scenario and only shows us a new scene. The scene scenario is written in an XML-like script. The video browser has a function to give effect according to operations of the scenario. In addition, our video viewing browser can provide us with an interactive video art. For example, suppose that a small stream runs down among the rocks. On the browser, if you chose an icon which shows maple leafs and drop it into the stream, a maple leaf starts floating down along the stream.

The authors developed a prototype indication display system for route guidance. We suppose that this system is set up in the public facilities especially a library. A city public library has a collection of many books. It is difficult for a visitor to find the desired book from many books. But we can recently use a library search system for inquiring the whereabouts of a book. The search system outputs a receipt which shows the search result. The destination is shown in this receipt. Though the visitor carries about the receipt during the treasure hunting, that receipt will be a trash after the arrival at his/her destination. The motivation of this research is to build the paperless whereabouts guide system. The guidance system using an electronic display would be possible to provide us with route information to the desired book.

The authors developed the handy route guidance system using a laser illuminated direction indicator. This system is applicable to the route guidance anywhere. This route guide system consists of a GPS receiver, a digital compass, a laser illumination and a portable computer for system integration. To indicate the direction, the navigation system needs to study a position on the earth. The information about a direction and a location is obtained by the GPS receiver and the digital compass. The laser illumination directly projects the direction indicator on the ground. The IC processor compares which the direction of a sensor is the same where a user should go or not while the user turns one revolution. Then the handy navigator lights a laser illuminator which shows the indicator on the ground if the direction is the same. Thus our navigation unit requests users to do exercise for an interaction to the earth.

Blind deconvolution is a significant technology in the restoration of atmospheric turbulence-degraded images. However, if the atmospheric turbulence-degraded images are contaminated by noise, the restoration images will be beyond real image due to involving a mount of noise. A novel blind deconvolution method has been proposed. In this method, the degraded image is preprocessed by a linear filter for reducing noise, and the filter is considered in the cost function of blind deconvolution. An alternating minimization algorithm based on conjugate gradient method is applied for minimizing the cost function. Thus, the smoothness induced by linear filter and the blur induced by atmospheric turbulence are eliminated in blind deconvolution simultaneously. For verifying this method, the images degraded by turbulence with atmospheric seeing parameter equal to 0.1 meters for 2 meters telescope and contaminated by noise with signal noise ratio equal to 10 dB are simulated by computer and restored by this method. The experiment result demonstrates that the noise is reduced without introducing any smoothing and the degraded image are restored effectively. The image restored by this method is compared with by the blind deconvolution method based on edge preserving regularization. The result shows that the effect of reducing noise of our method is better than the latter.

The remote camera developed by us is the exclusive functional load of the micro-satellite. The remote camera is based on the frame transfer CCD sensor DALSA FT18, and for the purpose of insuring system reliability, the development of the remote camera indispensably simplifies the design of mechanical and electrical shutter, which causes the problem of CCD smearing in remote sensors, and leads to the distortion of remote sensing images. In this paper we present a reversely stepwise method to solve the CCD smearing problem in remote sensors. The images retrieved from data after correction show great improvement in image contrast and quality.

An X-ray nondestructive detector for high-speed running conveyor belt with steel wire ropes is researched in the paper. The principle of X-ray nondestructive testing (NDT) is analyzed, the general scheme of the X-ray nondestructive testing system is proposed, and the nondestructive detector for high-speed running conveyor belt with steel wire ropes is developed. The hardware of system is designed with Xilinx's VIRTEX-4 FPGA that embeds PowerPC and MAC IP core, and its network communication software based on TCP/IP protocol is programmed by loading LwIP to PowerPC. The nondestructive testing of high-speed conveyor belt with steel wire ropes and network transfer function are implemented. It is a strong real-time system with rapid scanning speed, high reliability and remotely nondestructive testing function. The nondestructive detector can be applied to the detection of product line in industry.

Tracking by foreground matching heavily depends on the appearance model to establish object correspondences among frames and essentially, the appearance model should encode both the difference part between the object and background to guarantee the robustness and the stable part to ensure tracking consistency. This paper provides a solution for online maintaining appearance models by adjusting features in the model. Object appearance is co-modeled by a subset of Haar features selected from the over-complete feature dictionary which encodes the discriminative part of object appearance and the color histogram which describes the stable appearance. During the particle filtering process, feature values both from background patches and object observations are sampled efficiently by the aid of "foreground" and "background" particles respectively. Based on these sampled values, top-ranked discriminative features are added and invalid features are removed out to ensure the object being distinguishable from current background according to the evolving appearance model. The tracker based on this online appearance model maintaining technique has been tested on people and car tracking tasks and promising experimental results are obtained.

A novel method for image segmentation using double-level parallelized firing pulse coupled neural networks (DLPFPCNN) is presented in this paper. The first level (or auxiliary level) is used to enhance image by improved and simplified PCNN model combining with boundary enhancement, which can give the better results for the second level (or primary level) PCNN. The primary level uses a parallelized firing PCNN (PFPCNN) model to segment the enhanced images so that can improve the adaptability to the complex environment. Parallelized firing neuron model can overcome the drawbacks for sequential pulse burst, which is unfair for those pixels at low grayscale value areas. Finally, the optimal segmentation results are determined by maximum Shannon entropy of image. Experimental results show, as compared to the conventional PCNN model with single level and sequential pulse burst, the proposed method can improve the performance of image segmentation and obtain the good results, especially suiting for those images with low contrast, low signal-to-noise ratio (SNR) and continuously spatial-varying background.

Image registration is a process of creating correspondence between a pair of images. In some situations, the physical one-to-one correspondence may not exist due to the presence of "outlier" objects (called gross outliers) that appear in one image but not the other. In this paper, a novel robust method is presented to address the problem of tumor-like gross outliers in non-rigid image registration. First, two salient point sets are extracted from the two images to be registered, and classified by means of clustering analysis which is based on Gaussian mixture models and expectation-maximization (EM) algorithm. Then by means of joint saliency map that represents the joint salient regions of the overlapping volume of the two images, the regions including tumor-like gross outliers could be automatically recognized. After screening out of salient points and elimination of outlier points, some stable control points that well represent the corresponding structures within the joint salient regions of the two images could be obtained. By iteratively finding correspondences between the control points in the joint salient regions, the smooth deformation field is approximated based on radial basis functions (RBFs) with compact support until the convergence to the steady-state solution is achieved. Experimental results show that the proposed method is able to recover local deformation caused by tumor resection in brain.

Multimedia technology and networks protocol are the basic technology of the video surveillance system. A network remote video surveillance system based on MPEG-4 video coding standards is designed and implemented in this paper. The advantages of the MPEG-4 are analyzed in detail in the surveillance field, and then the real-time protocol and real-time control protocol (RTP/RTCP) are chosen as the networks transmission protocol. The whole system includes video coding control module, playing back module, network transmission module and network receiver module The scheme of management, control and storage about video data are discussed. The DirectShow technology is used to playback video data. The transmission scheme of digital video processing in networks, RTP packaging of MPEG-4 video stream is discussed. The receiver scheme of video date and mechanism of buffer are discussed. The most of the functions are archived by software, except that the video coding control module is achieved by hardware. The experiment results show that it provides good video quality and has the real-time performance. This system can be applied into wide fields.

In this paper, a method of IFS-based image geometry transform is proposed. Suppose the original image can be approximated with the attractor (denoted by A) of an Iterated Function System (IFS) consisting of N contractive mappings of w_n (n=1, 2, ..., N), whose coefficients have been determined by fractal encoding. G(A) is used to denote the geometry transform on the attractor A. The result is equivalent to make a corresponding geometry transform on the original image. It is demonstrated in the paper that G(A) is the attractor of a new iterated function system (denoted by IFS') derived from the mappings of w_n (n = 1, 2, ...N). In another word, we can modify the coefficients of w_n (n=1, ..., N) to construct the IFS', and the result by decoding IFS' is A' = G(A), which is the approximation of the expected geometry transform of the original image. In order to translate, rotate and dilate images in the domain of IFS coefficients, formulas to construct the IFS' from w_n are deduced in this paper. The experimental results have validated the proposed method.

Spatiotemporal patterns are obtained in gas discharge with increasing the voltage with the evolvement as following: random spots, quasihexagonal pattern, hexagonal pattern with large spot, hexagonal pattern with small spot, white-eye pattern, and chaos state. The structures of patterns are studied by analyzing the spatial correlation functions, Fourier spectrum and angular spectral distributions of patterns after the image processing to the images of patterns. Furthermore, the mechanism of the pattern formation is investigated, which indicates that hexagonal patterns both large and small spot are selected by the single mode, while the white-eye pattern results from a three wave resonance.

Some limits of standard Kalman filtering are simply analyzed. Such as the indefiniteness of motion resulted from targets maneuvering and the lower predictive precision brought about by non or little adaptive capabilities make standard Kalman filtering lower tracking precision and stabilization. Interacting multiple model algorithm is adopted to combine with Kalman filter, and a new adaptive Kalman filtering algorithm for improving tracking capabilities is proposed. Multiple models are designed to represent system possible running patterns, and "current" statistical model is designated as one of them. Each model has an independent Kalman filter, and the general state estimation is a kind of mixing data output produced by interacting among these models' state estimations through certain mixing probabilities. Each model state estimation is produced by one Kalman filter corresponding this system model. In simulation tests, three system models are designed to work, CV model , CA model and "current" statistic model. Tests show that the indefiniteness resulted from the target motional model approximately describing the target motional pattern, the lower adaptive tracking capabilities, and the lower tracking precision and stabilization of in targets tracking are improved efficiently. Moreover, the strong nonlinear problem is solved effectively.

The IWT (Integer Wavelet Transform) can achieve genuine lossless image compression and allow both lossy and lossless compression using a single bit-stream. However, using the IWT instead of the DWT (Discrete Wavelet Transform) will degrade the performances of the lossy compression because the filter structure of IWT and the nonlinear rounding operation. In this paper, a new integer wavelet decomposition scheme is proposed based on subband local information statistic model for the medical images. The high frequency subbands can be decomposed again according to the statistic results of the subband coefficient entropies. The results of several experiments for the medical images presented in this paper demonstrate the importance subband information statistic in the integer wavelet decomposition. Furthermore, this paper shows that appropriate subband local information statistic model improves the performance of compression algorithm after the multilevel subband decomposition is performed. So we expect this idea is valuable for future research on medical image coding.

For low bit rate video compression, the quality of reconstructed video is usually poor. The high codec priority of region of interest (ROI) can improve image quality obviously. Nowadays, video segmentation methods are often used for extracting ROI, but these methods have high computational complexity and are not satisfied to real time communication. On the other hand, in most existing rate control algorithms, ROI can't select the low and high bit rate R-Q model adaptively. Aiming at these problems, in this paper, a simple and efficient approach of extracting ROI is proposed which can decrease the computational complexity of existing ROI extracting algorithms. Bits are distributed to ROI and non- ROI (NROI) respectively according to the image complexity and motion information. Moreover, the judgment criterion of distinguishing between low and high bit rate coding category is derived, which makes the algorithm select the R-Q model adaptively and decrease the rate control errors. In addition, the scheme of modifying the coding order of macro blocks (MBs) can enhance the objective image quality. Experiment results demonstrate that the proposed algorithm achieves a bit rate closer to the target, provides fewer skipped frames, and gets better objective and subjective image quality significantly compared with TMN7 and TMN8 algorithms.

A improved SPIHT image compression algorithm called symbol-map zero-tree coding algorithm (SMZTC) is proposed in this paper based on wavelet transform. The SPIHT algorithm is a high efficiency wavelet coefficients coding method and have good image compressing effect, but it has more complexity and need too much memory. The algorithm presented in this paper utilizes two small symbol-maps Mark and FC to store the status of coefficients and zero tree sets during coding procedure so as to reduce the memory requirement. By this strategy, the memory cost is reduced distinctly as well as the scanning speed of coefficients is improved. Those comparison experiments for 512 by 512 images are done with some other zerotree coding algorithms, such as SPIHT, NLS method. During the experiments, the biorthogonal 9/7 lifting wavelet transform is used to image transform. The results of coding experiments show that this algorithm speed of codec is improved significantly, and compression-ratio is almost uniformed with SPIHT algorithm.

The novel visible nephogram imaging technology for polar orbit platform is demonstrated in the paper, and it could be operated in from quarter moon to noon sunlight. The critical technologies and theirs solutions of the novel nephograph are included: (i) the low light level imaging capability is achieved by the combination of time delay and integration charge coupled device (TDI CCD) with push-broom imaging method; (ii) the large field of view capability is implemented by the combination of 3 pieces of imaging module with smaller field of view; (iii) the wide dynamic range capability is achieved by the combination of TDI CCD with gradient neutral density filter (NDF). On the basis of the analysis and trade-off of system design, the prototype of novel visible nephograph for polar orbit platform is developed. The results of experiments and tests in ground demonstration are satisfying, and the nephograph prototype is mainly met the customer demand. In the end of paper, several problems and theirs solution of novel technology for space application are also mentioned.

A novel content-based image retrieval data structure is developed in present work. It can improve the searching efficiency significantly. All images are organized into a tree, in which every node is comprised of images with similar features. Images in a children node have more similarity (less variance) within themselves in relative to its parent. It means that every node is a cluster and each of its children nodes is a sub-cluster. Information contained in a node includes not only the number of images, but also the center and the variance of these images. Upon the addition of new images, the tree structure is capable of dynamically changing to ensure the minimization of total variance of the tree. Subsequently, a heuristic method has been designed to retrieve the information from this tree. Given a sample image, the probability of a tree node that contains the similar images is computed using the center of the node and its variance. If the probability is higher than a certain threshold, this node will be recursively checked to locate the similar images. So will its children nodes if their probability is also higher than that threshold. If no sufficient similar images were founded, a reduced threshold value would be adopted to initiate a new seeking from the root node. The search terminates when it found sufficient similar images or the threshold value is too low to give meaningful sense. Experiments have shown that the proposed dynamic cluster tree is able to improve the searching efficiency notably.

A new detection algorithm of dim moving targets in the IR image sequences is presented. The images consist mainly of sensor noise, drifting background clutter and low contrast targets. So it is difficult to provide reliable detection in just a single frame. This algorithm adopts multi-scale gradient to suppress the clutter according to its spatial distribution feature. Then the recursive maximum similarity (RMS) filter is used to accumulate targets energy in temporal domain. The advantage of the algorithm is that it realizes the enhancement of the target gradient feature and clutter suppressing in the same time. The results show that the algorithm can effectively detect the real small dim targets even if there is strong clutter influence.

Using the local modeling method can get sub-pixel localization in digital image measurement, in order to improve the precision of sub-pixel measurement which is usually obtained by computing the properties of fitting curves in local modeling method, an effective method is presented. The proposed method is that at first using bilinear interpolation or other more reasonable interpolation ways to original image obtains a new image and then some edge features are extracted from the new image by chain code edge detection or other ways. Due to using of interpolation, after the pixels of extracted edge features are mapped to original image size, the edge features will contain more pixels which belong to integer pixels from original image and the sub-pixels which the interpolation image produces, subsequently those pixels will be used to fit curve to improve the accuracy of the fitting. According to the result of the improving fitting, a more accurate measurement is obtained by utilizing the properties of the curves. At the last section of this paper, the proposed method is evaluated by using real images which is collected by digital camera, the experiment result turns out that the algorithm owns better accuracy than the one which is only by only fitting on account of the proposed method which owns better standard deviation than the only fitting way without interpolation.

Edge detection is realized by using a FitzHugh Nagumo model which is one type of partial differential equations. This model has three types of dynamic, excitable, Turing/Hopf bifurcation, and bistable. In the excitable region the model can realize the edge detection. In the simulation only one image is processed in order to confirm the effect of control parameters on the edge detection. A satisfying effect of edge detection can be obtained by choosing appropriate control parameters. By comparing with other operators it is found that the FitzHugh Nagumo model is superior to Canny operator, Prewitt operator, Roberts operator, and Sobel operator on edge detection.

A novel method for automated defect detection of pearls based on machine vision is proposed. Firstly, a dome-shaped light source with diffused light illumination was designed to improve image quality and reduce light-spot size. And a novel quasi-synchronous multi-images grabbing scheme from different views is then designed based on pearl' free-falling motion. Then a nonlinear filter based on space geometry is given to enhance defect contrasts following by a region-grow method for extracting all suspicious defects, including highlight-halation regions. Furthermore, the highlight-halation regions were removed using morphological method based on the spatial distributive model of the highlight-halation. At last, shape and texture features of defect regions are extracted and SVM method was used for defect grading. Experiments show that the acquired images included the complete information of pearl surfaces and the system correctness was over 93.3% .

To improve the retrieval accuracy of content-based video retrieval systems, researchers face a hard challenge that is reducing the 'semantic gap' between the extracted features of the systems and the richness of human semantics. This paper presents a novel video retrieval system to bridge the semantic gap. Firstly, the video captions are segmented from the video and then are transformed into text format. To extract the semantic information from the video streaming we apply a text mining process, which adopts a cluster algorithm as a kernel, on the text format captions. On the other hand, in this system, users are requested to comment on the video which they download from the system when they have watched the video. Then we associate the users' comments with the video on the system. The same text mining process is used to deal with the comment texts. We combine the captions of the video with the comments on the video to extract the semantic information of the video more accurately. Finally, taking advantage of the comments and the captions of the video, we performed experiments on a set of videos and obtained promising results.

Visual inspection is a kind of new technology based on the research of computer vision, which focuses on the measurement of the object's geometry and location. It can be widely used in online measurement, and other real-time measurement process. Because of the defects of the traditional visual inspection, a new visual detection mode -all-digital intelligent acquisition and transmission is presented. The image processing, including filtering, image compression, binarization, edge detection and distortion correction, can be completed in the programmable devices -FPGA. As the wide-field angle lens is adopted in the system, the output images have serious distortion. Limited by the calculating speed of computer, software can only correct the distortion of static images but not the distortion of dynamic images. To reach the real-time need, we design a distortion correction system based on FPGA. The method of hardware distortion correction is that the spatial correction data are calculated first under software circumstance, then converted into the address of hardware storage and stored in the hardware look-up table, through which data can be read out to correct gray level. The major benefit using FPGA is that the same circuit can be used for other circularly symmetric wide-angle lenses without being modified.

Mercuric iodide (HgI₂) polycrystalline films are being developed as a new photoconductor layer for direct converter X-ray imaging detector. A physical vapor deposition (PVD) device for HgI₂ deposition was developed specially. Depending on the device and low purity (99.5%) low-cost HgI₂ source material, polycrystalline HgI₂ films have been grown with dimensions Φ130 mm in diameter onto ITO-coated glass substrate. The grown techniques used can be easily extended to produce much larger films areas and the thickness of the grown layers, size of the grains and crystallinity can be regulated in a controlled way by adjusting the growth parameters. The basic physical characteristic, dark current and response characteristic to X-ray for the grown polycrystalline film were tested and the results show that the film has preferred crystalline orientation (00l), low dark current density less than 10 pA/mm², high volume resistivity in the order of 10¹³ Ω•cm and high X-ray response sensitivity of about 16 μC/(cm²•R), and the results put HgI₂ polycrystalline films in position as a leading candidate material for use in digital X-ray imaging system.

Contrast enhancement is important for small target detection and tracking. Conventional contrast enhancement techniques often fail to produce satisfactory results for images expressing unimodal intensity histograms. This paper presents a new contrast enhancement method based on rough set theory which is especially suitable for such images. The method uses the max of between-class mean to partition the image into two sub-images, the denoised target region and the denoised background region. Then the target region is enhanced by extend histogram. Experimental results indicate the new enhancement method is more suitable than traditional methods for handling the enhancement problems of low contrast small target images.

In this paper, how to achieve 3D visualization fault diagnosis system for photoelectric tracking equipment based on open graphic library(OpenGL) is researched. To begin with, details of the system architecture design and implementation are presented. The 3D modelings of all the equipments are built by using 3DSMAX software. Then, the model is transformed into OpenGL programs. This method overcomes the difficulty of building complex model directly using OpenGL and reduces the modeling workload. While implementing 3D driving, the alternative operation between human and the computer is achieved. Finally, intelligent fault diagnosis technologies including the rule base and the reasoning strategy in the expert system are discussed. Practical applications illuminate that the proposed approach is feasible and effective.

Frame transfer area-array CCD camera is the perfect solution for high-end real-time medical, scientific and industrial applications because it has characteristics of high fill factor, low dark current, high resolving power, high sensitivity, high linear dynamic range and electronic shutter capability. Time sequences of frame transfer area-array CCD camera have two compact segments: CCD driving sequences and CCD signal processing sequences. Proper working of CCD sensor lies on good driving sequences while accurate CCD signal processing sequences ensures high quality of CCD image. The relationship among CCD camera time sequences is complex and precise. The conventional methods are uneasy to implement time sequences of Frame transfer area-array CCD. Embedded designing method is introduced in this paper and field programmable gate array device is chosen as the hardware design platform. Phase-locked loops are used for precise phase shifting and embedded logic analyzer for waveform verification. CCD driving clocks, electronic shutter signal, A/D and black pixels clamp clocks and double correlation sampling clocks have been attained on the hardware platform and this timing generator can control exposure time flexibly. High quality images have been acquired through using this timing generator on the CCD circuit system board which has been designed by our team.

The measuring technology of minute part's geometrical parameter based on image processing is an integration of optics, the mechanics, electronics, calculation and control. Accomplishing the video alteration of measuring microscope, real-time gathering image with CCD, and compiling automatically measuring software in Visual C++6.0 environment. First to do image processing which includes denoise filter, illuminance non-uniformity adjustment and image enhancement, then to carry on the on-line automatic measuring to its geometry parameters. By measuring the minute part's geometry parameters of machineries and integrated circuit in this system, the experimental results indicate that the measuring accuracy could amount to 1 micron, and the system survey stability and usability are all good.

In order to improve the efficiency and to supply more sufficient information support, an intelligent fault diagnosis system based on desktop virtual environment is proposed. In the first place, basic concepts and principles of virtual reality and intelligent fault diagnosis technology are presented in this paper. Then, several essential implementation issues of the system, including the system architecture, the 3D visualization of the fault diagnosis environment and the user interface and so on, are also been discussed. Lastly, intelligent fault diagnosis technologies are elaborated, such as the rule base and the strategy of the reasoning and control in the expert system, etc. Practical applications and experiments demonstrate that the proposed approach is effective and robust.

In this paper, a novel contourlet transform based on median filter is proposed. By using a novel median pyramidal decomposition, the noise distributing is analyzed for the image distorted by salt-and-pepper noise and Gaussian noise respectively. Comparing the Probability-Density -Functions of the detail coefficients of the each corresponding layer, it is found that these two kinds of noise mainly concentrate on the bottom high frequency layer. So a majority of noises can be removed by denoting zero the bottom layer coefficient. Median-Contourlet transform is completed when the second layer and other high frequency image is calculated by PDFB(Pyramidal Directional Filter Bank). After analysing of Contourlet coefficients, we select the best threshold to remove further the noises. Applying the same denoising method to images, the Median-Contourlet achieves obvious improvement in both subjective visual effect and SNR comparing with traditional contourlet transform.

In the industrial and network monitoring field, there are several protocols such as Pelco D/P used for remote operation and widely applied to control the pan/tilt/zoom (PTZ) camera systems. But for universal CCD camera, a lot of incompatible communication protocols have be developed by different manufacturers. To extend these cameras' application in remote monitoring field and improving its compatibility with controlling terminal, it's necessary to design a reliable protocols conversion module. This paper aimed at realizing the conversion and recognize of different protocols for CCD camera. Protocol conversion principle and algorithm are analyzed to implement instruction transformation for any camera protocols. An example is demonstrated by converting Protocol Pelco D/P into Protocol 54G30 using Micro Controller Unit (MCU). High performance hardware and rapid software algorithm was designed for high efficient conversion process. By means of serial communication assistant, Video Server and PTZ controlling keyboard, the stability and reliability of this module were finally validated.

With the fast-speed real-time wavefront processor in 37-element adaptive optical (AO) system as example, the limitation of conventional wavefront processor is analyzed, a novel wavefront processor is proposed based on ARM-based embedded system. The method will enable the whole 37-element AO system compacter, smarter and more effective. The merit of this processor will make the AO system be suitable for more especial situation. The hardware configuration and the method of software based on Linux operating system are both exposited in detail.

The influence of different water temperatures on temporal behavior of dielectric barrier discharge in argon at atmospheric pressure is studied by using an experimental device with water electrodes. It is found that, as the applied voltage increasing, the evolvement of discharge with different water temperatures shows similar behavior spatially if argon concentrations are same. The corresponding Fourier transformation is obtained by processing patterns images with computer program based on the Matlab software. The discharge duration of the first current pulse in half cycle of different voltage polarity is same in discharge if the two electrodes temperatures are same. But it becomes different if the two electrodes have different temperatures. The discharge moment is always ahead when the low temperature electrode is an instantaneous cathode. The analysis shows that the water temperature affects the accumulation of wall charges, resulting in the differences of temporal behavior of discharge.

Object-identification using edge extraction techniques from background function in uncontrolled lighting environments containing more object pose information and have many applications. In order to depressing noise, identify aim body robustly and rapidly, in this paper, We take cuboid as model and present a new strategy for edge extraction and object-identification based on object inherent features. This strategy includes the following steps. Firstly, pre-processing is applied to the raw image, in which Canny operator was used to extract edges pixels, then, image was divided into a grid of overlapping windows and noise was suppressed by regression grid windows in which the number of pixels is less than a threshold. Secondly, as model contour's geometry characters known already, the cuboids upright edges was used as their existence evidence to estimate model's existence area and so the lines failed spatial constraints are eliminated, then, object edges was extracted within the finite ranges of orientation in Hough transform space. Thirdly, the intersections of the component extracted edges are taken, the candidate edges extraction and matches was assessed based on the intersections, rather than the component extracted edges. After a series of matching tests the aim body is extracted. The proposed method makes three major contributions. Firstly, on the base of study the correspondence between model's boundary edges parameters in image space and Hough space we extract edges in finite area in Hough transform space, the aimless computations and searching is reduced greatly, its efficiency improved. Secondly, as Canny operator can extract aim lines with single pixel width, the edges extraction strategy of combining Canny operator with Hough transform extractor could avoid error impact of edges pixels numbers to Hough extractor. Thirdly, after fusion model's knowledge in image space, Hough space, global space, learning from others strong points to offset one's weakness, we extract model's edges from complex noise background without regarding to regression caused by the errors due to spurious or missing pixels because edge extraction is imperfect for real images. The results of experiments demonstrated that the proposed method could suppress noise effectively, identified and extracted target from complex backgrounds robustly. This new strategy may have potential application in visual servo, object tracking, port AGV and robots fields etc.

Image enlargement technique is widely used in image processing and video format transforming. This paper proposes an algorithm for fractional multiple image enlargement based on All Phase Discrete Cosine Transform (APDCT) interpolation. The All Phase Digital Filter (APDF) is a new type of linear phase filter. For a data vector with a length of N obtained by blocking a signal, there are N different phase data blocks that include the same sampling point. Through taking the mean of the N values as the filtering output, the APDF can eliminate different meanings of the orthogonal transform filtering values of those data vectors and thereby the block effect. According to this idea, this paper deduces the formula of two-dimension APDCT filter. As a good integration of Discrete Cosine Transform (DCT) and all phase idea, a family of separable interpolation kernel, which is constructed by cosine function, is put forward. Look-up method is then discussed for real time image enlargement. The objective and subjective experiments show that better image quality is obtained by using interpolation kernel proposed than using 6*6 cubic interpolation kernel.

In the evaluation of surface roughness by computer vision technique, the pattern of illumination is generally correlated with optical surface finish parameters from the images. So this paper carried out experiments to investigate the effects of various factors and completed the optimum design of capture condition. Then we captured abundant sample images under appropriate experimental condition and chose to extract features of surface roughness in the spatial frequency domain which should be less sensitive to noise than spatial domain features. Therefore, artificial neural network (ANN), which took frequency-domain roughness features as the input, was developed to determine surface roughness by selecting the back-propagation algorithm. The built ANNs using these critical sets of inputs showed low deviation from the training data, low deviation from the testing data and high sensibility to the inputs levels. And the high prediction accuracy of the developed ANNs was confirmed by the good agreement with the results from traditional stylus method. Hence the proposed roughness features and neural network were efficient and effective for automated assessment of surface roughness.

Land surface vegetation phenology is an important process for the real-time monitoring and detecting inter-annual variability in terrestrial ecosystem carbon exchange and climate-biosphere interactions. Crop phenology is an important factor that influences crop growth and yield estimation models. Since the mid-1980s, coarse-resolution, temporally-composited satellite data have been used to study vegetation phenology. View-angle corrected nadir reflectances from the 16-day, 1km operational MODIS BRDF/Albedo product are currently used to monitor global land cover dynamics. In this paper, we developed an improved methodology for using the new 500-m MODIS BRDF/Albedo Version 005 product to monitor global vegetation phenology by utilizing time series of the Normalized Difference Vegetation Index (NDVI). The method adopts a rolling strategy for the continuous updating of the underlying anisotropy (or BRDF shape), so that the latest land surface BRDF information can be used as prior-knowledge for next retrieval. Using this approach, transition dates for vegetation phenology in time series of NDVI can be determined from MODIS data at finer temporal and spatial resolution. Preliminary results based on monitoring crops in northern China demonstrate the effectiveness of our rolling retrievals coupled with the improved spatial resolution of the new MODIS product.

A new approach to recover 3D structure from the uncalibrated image sequence is presented. Unlike previous methods, the method recovers 3D models' initial values by enforcing the geometric constraints of the straight line in the scene and then gets accurate results by bundle adjustment. The focal length and the principal point of each camera are supposed to be unknown intrinsic parameters. The similarity invariance of the straight line that vertical straight line pairs in the scene keep vertical in similarity transformation are taken account of as a geometric constraint. A projective reconstruction performed by an iterative factorization algorithm is upgraded to a Euclidean one with the constraint. Bundle adjustment refines a visual reconstruction to produce jointly optimal structure and viewing parameter estimates when we get the ideal initial values from the method above. Experiments results on synthetic and real image sequences verified the new approach's precise and efficiency.

To meet the requirements of large-scale hyperspectral image analysis and identification applications, a processing flow of georeferenced mosaic is set up particularly for Hyperion true color images with overlapping areas. The method mainly includes algorithms of brightness balancing and cutline feathering. Advanced weight-smoothing is applied to blend image boundaries as a feathering technique, and brightness balancing is fulfilled using Improved-Compensation and Histogram Matching. In these improved methods gray threshold is setting extra for features of Hyperion data, which include random high brightness factor (e.g. cloud and mist). Furthermore, all these methods are easily operated and quite effective for massive data such as Hypeiron images. Finally, a mosaic thematic map (with a scale of 1:100,000) of 8 scenes of Hyperion images is produced based on the research in this paper, which makes the image boundaries natural and provides a good visual result.

Image fusion could process and utilize the source images, with complementing different image information, to achieve the more objective and essential understanding of the identical object. Recently, image fusion has been extensively applied in many fields such as medical imaging, micro photographic imaging, remote sensing, and computer vision as well as robot. There are various methods have been proposed in the past years, such as pyramid decomposition and wavelet transform algorithm. As for wavelet transform algorithm, due to the virtue of its multi-resolution, wavelet transform has been applied in image processing successfully. Another advantage of wavelet transform is that it can be much more easily realized in hardware, because its data format is very simple, so it could save a lot of resources, besides, to some extent, it can solve the real-time problem of huge-data image fusion. However, as the orthogonal filter of wavelet transform doesn't have the characteristics of linear phase, the phase distortion will lead to the distortion of the image edge. To make up for this shortcoming, the biorthogonal wavelet is introduced here. So, a novel image fusion scheme based on biorthogonal wavelet decomposition is presented in this paper. As for the low-frequency and high-frequency wavelet decomposition coefficients, the local-area-energy-weighted-coefficient fusion rule is adopted and different thresholds of low-frequency and high-frequency are set. Based on biorthogonal wavelet transform and traditional pyramid decomposition algorithm, an MMW image and a visible image are fused in the experiment. Compared with the traditional pyramid decomposition, the fusion scheme based biorthogonal wavelet is more capable to retain and pick up image information, and make up the distortion of image edge. So, it has a wide application potential.

Based on the research of the standard second generation, the high capability third generation, the exceeding third generation and the fourth generation, the spectral response of the third generation LLL was carried out using a self-developing spectral response measurement instrument. Spectral response characteristics of third-generation LLL tube were obtained, and the material performance parameters of GaAs photocathode were calculated by curve simulation method. On-line measurement of GaAs photocathode after high-temperature activation and low-temperature activation was carried out, the results showed that spectral response in the whole response waveband decreased after indium seal, and long wave responsibility was most obviously influenced. Decrease was large, cut-off wavelength and peak value wavelength move towards short-wave, peak response value and integral sensitivity decreased, and the final spectral response curve became flat. By calculating photocathode parameters, it was found that indium seal, lead to the variations of surface activation layers of photocathode, and the long wave responded and sensitivity decreased accordingly. The influence factors on the surface activation layers during indium seal were also analyzed.

With the development of human computer interaction, lip reading technology has become a topic focus in the multimode technologic field. However, detecting and locating lip accurately are very difficult because lip contours of different people, varied illuminant conditions, head movements and other factors. Based on the methods of detecting and locating lip we proposed the methods which are based on the lips color extracted lip contour using the adaptive chromatic filter from the facial images. It is not sensitive to illumination, but appropriate chromatic lip filter is given by analyzing the entire face color and clustering statistics of lip color. It is proposed the combinable method which is preprocessing the face image including rotating the angle of face and improving image contrast in this paper and the lip region is analyzed clustering characteristics for the skin color and lip color, obtained adaptive chromatic filter which can prominent lips from the facial image. This method overcomes the varied illuminate, incline face. The experiments showed that it enhanced detection and location accurately through rough detecting lip region. It lay a good foundation for extraction the lip feature and tracking lip subsequently.

With the development of the micro-processing, silicon material technique, the silicon structural design, mass production, the scientific researches on integrated circuit have been developed rapidly. Defect detection and fault diagnosis as critical design requirements is necessary to achieve high-quality, cost-effective multichip systems. An increasingly difficult task, however, is the inherent need to accurately locate, indentify the defects within the microchip. A method of defect detection of IC wafer have been investigated using mathematical morphology. Firstly, the differential charts of the pending images and the intact image of IC wafer are computed and digitized to two gray levels, i.e. black and white. Secondly, the brightness of the pending images have been transformed to the same brightness as that of intact image in order to make the arithmetics is robust for various illumination conditions. Next, the defects on the binary image of differential chart can be found and dealt with mathematical morphology. Finally, several representative characteristics are proposed to extract and describe the defects of IC wafer image, for example perimeter, area, macro axis, minor axis, eccentricity ratio, centroid, circularity and rectangular degree, etc.

Embedded block coding with optimized truncation (EBCOT) with the wavelet shape of tree encoding structure is more flexible because it encodes each code block respectively by decomposing the subband into code blocks, so that the embedded code streams will come into being to support the mass classification, the hierarchical resolution and the random access. However the anti-missing performance via network of the algorithm is worse. The source signal has been divided into many code streams by multi-description coding (MDC) of image and video so that it will be transferred through the insecure transmission channel. Region of Interest (ROI) coding gives priority to the focus of doctor's interest generally occupies lesser part of entire medical image. In this paper, ROI coding, which combines MDC and EBCOT, has been done according to JPEG2000 ROI coding standard in medical images. The algorithm not only uses the hierarchical spatial resolution and random access to ROI of EBCOT, but also has improved the anti-missing performance via network, and formed robust code stream. The experimental results demonstrated that the coding method improved the system compression ratio without influence on the medical diagnosis.

Content-based multimedia information retrieval is an interesting but difficult area of research. Current approaches include the use of color, texture, and shape information. This paper proposes a novel approach to content-based image retrieval. In our retrieval system, an image is represented by a set of color histogram and edge histogram descriptors. The histogram Euclidean distance, cosine distance and histogram intersection are used to measure the image level similarity. Finally, the overall similarity is computed as a weighted combination of image similarity measures incorporating all features. Our proposed retrieval approach demonstrates a promising performance for an image database including 766 general-purpose images. Effectiveness is documented by experimental results.

With high-speed development of information technology, the demand for multi-media image processing has become more and more urgent. According to characters of embedded system and video image, the paper proposes an embedded system that can recognize the moving objects in video stream, and give the realization of hardware and software. Experiment result shows that system have a good effect in better light.

This paper describes a high-frame rate CCD camera capable of operating at 100 frames/s. This camera utilizes Kodak KAI-0340, an interline transfer CCD with 640(vertical)×480(horizontal) pixels. Two output ports are used to read out CCD data and pixel rates approaching 30 MHz. Because of its reduced effective opacity of vertical charge transfer registers, interline transfer CCD can cause undesired image artifacts, such as random white spots and smear generated in the registers. To increase frame rate, a kind of speed-up structure has been incorporated inside KAI-0340, then it is vulnerable to a vertical stripe effect. The phenomena which mentioned above may severely impair the image quality. To solve these problems, some electronic methods of eliminating these artifacts are adopted. Special clocking mode can dump the unwanted charge quickly, then the fast readout of the images, cleared of smear, follows immediately. Amplifier is used to sense and correct delay mismatch between the dual phase vertical clock pulses, the transition edges become close to coincident, so vertical stripes disappear. Results obtained with the CCD camera are shown.

A novel parallel image processing architecture using multiple DSPs which can satisfy real-time image processing demands is proposed, The architecture is structured with high performance DSP interconnected by FPGA. Within FPGA the interconnection network by IRAM and the specific data communication protocol are implemented. The system inherits merits from the tightly coupled parallel system and the loosely coupled parallel system. The system architecture is reconfigurable and scalable. The performances measured in this platform show the high data transfer rate, and it can satisfy parallel real-time image processing demands of the complex task, large computation and high-speed data transfer. From the designed parallel hardware we analyze the benchmarks including acceleration ratio, parallel efficiency, selection of processing units, interconnection network etc. Finally some suggestions are given to further improve the system performance. The real-time image processing system based on parallel structure of multiple DSPs is easy to be implemented. Because the system structure is reconfigurable and scalable, it is easy to change the number of DSP and change the DSP into other series. So it has a bright future for the application of real-time image processing system.

Sand, chemical waste, microbes and other solid materials flowing with the water bodies are of great significance to us as they cause substantial impact to different sectors including drinking water management, hydropower generation, irrigation, aquatic life preservation and various other socio-ecological factors. Such particles can't completely be avoided due to the high cost of construction and maintenance of the waste-treatment methods. A detailed understanding of solid particles in surface water system can have benefit in effective, economic, environmental and social management of water resources. This paper describes an automated system of fluvial particle characterization based on spectral image processing that lead to the development of devices for monitoring flowing particles in river. Previous research in coherent field has shown that it is possible to automatically classify shapes and sizes of solid particles ranging from 300-400 μm using artificial neural networks (ANN) and image processing. Computer facilitated with hyper spectral and multi spectral images using ANN can further classify fluvial materials into organic, inorganic, biodegradable, bio non degradable and microbes. This makes the method attractive for real time monitoring of particles, sand and microorganism in water bodies at strategic locations. Continuous monitoring can be used to determine the effect of socio-economic activities in upstream rivers, or to monitor solid waste disposal from treatment plants and industries or to monitor erosive characteristic of sand and its contribution to degradation of efficiency of hydropower plant or to identify microorganism, calculate their population and study the impact of their presence. Such system can also be used to characterize fluvial particles for planning effective utilization of water resources in micro-mega hydropower plant, irrigation, aquatic life preservation etc.

Erosion of hydro turbine components through sand laden river is one of the biggest problems in Himalayas. Even with sediment trapping systems, complete removal of fine sediment from water is impossible and uneconomical; hence most of the turbine components in Himalayan Rivers are exposed to sand laden water and subject to erode. Pelton bucket which are being wildly used in different hydropower generation plant undergoes erosion on the continuous presence of sand particles in water. The subsequent erosion causes increase in splitter thickness, which is supposed to be theoretically zero. This increase in splitter thickness gives rise to back hitting of water followed by decrease in turbine efficiency. This paper describes the process of measurement of sharp edges like bucket tip using digital image processing. Image of each bucket is captured and allowed to run for 72 hours; sand concentration in water hitting the bucket is closely controlled and monitored. Later, the image of the test bucket is taken in the same condition. The process is repeated for 10 times. In this paper digital image processing which encompasses processes that performs image enhancement in both spatial and frequency domain. In addition, the processes that extract attributes from images, up to and including the measurement of splitter's tip. Processing of image has been done in MATLAB 6.5 platform. The result shows that quantitative measurement of edge erosion of sharp edges could accurately be detected and the erosion profile could be generated using image processing technique.