最新刊期
卷 8 ， 期 3 ， 2003
- 摘要：In order to improve the ability of real time computation, an attention-based IR ship detection is proposed to reduce the data to be processed and enhance the processing efficiency. The attention is divided into "pre-attention" and "attention". A nonlinear sampling model is adopted to reduce the resolution in pre-attention while keeping adaptive to size variance of the target; the hot region which always refers to the engine or chimney of a ship is adopted as the guidance of attention to the areas of interest(AOI), then the waterline, which is taken as the less salient feature of infrared ship target, is detected in the AOI. If the waterline feature exists in an AOI, it means a target is detected; otherwise, the AOI is taken as false alarm. To test the performance of the approach proposed, an algorithm is designed and realized both on PC and on a multi-DSPs system. Experiments demonstrate that the approach proposed can enhance the detection efficiency, and it is adaptive to the size of target.3622|192|0更新时间：2024-05-07
- 摘要：Pseudo-Zernike is a kind of region-based shape descriptor, contour-based shape descriptors, such as curvature descriptors, Fourier descriptors, signature or chain code-based descriptors, are not appropriate for describing shapes consisting of several disjoint regions, since they are often based on a single contour. There is no information redundancy because the bases of pseudo-Zernike moments are orthogonal. The magnitude of pseudo-Zernike moments has rotational invariant property. An image can be better described by a small set of its pseudo-Zernike moments than any other types of moments such as geometric moments, Legendre moments, rotational moments, and complex moments in terms of mean-square error. A relatively small set of pseudo-Zernike moments can characterize the global shape of a pattern effectively. Lower order moments represent the global shape of a pattern and higher order moments represent the detail. Pseudo-Zernike moments do not possess scale invariance. In this paper, we analyze the invariance of pseudo-Zernike moments and present an improved pseudo-Zernike moments that not only have rotation invariance, but also have scale invariance, and the same time, we give part of analysis result. The experimental results show that the improved Zernike moments have better invariant properties than unimproved Zernike moment using as region-based shape descriptor.3242|189|0更新时间：2024-05-07
- 摘要：According to the shortcoming of the classical gradient operator, this paper gives an improved computational formula for edge direction. The edge direction of the classical gradient operator is given by an anti-tangent function, and the angle range which the anti-tangent function can represent is [-π/2,π/2], but the range of edge direction is [0,2π]. This paper analyzes the operator from the gradient operators, and points out that the classical gradient operator can not discrete two edges which have a difference ofπin edge direction, and gives a improved computational formula for edge direction. In the classical gradient operator, because the computational formula for edge direction is not precise, the following image edge thinning, short edge removing, aperture filling and edge linking are constructed on a non-precision foundation, error results being got. Because of the error in low-level image processing, the quality of following high-level image processing is low. The computer-simulated experiments show that, the improved computational formula of edge direction in this paper is correct and effective.2938|208|0更新时间：2024-05-07
- 摘要：In the last decade, Active Contour Model(Snakes) has been successfully applied in the areas of computer vision and image processing. In particular, it reveal much potentials for edge detection and tracking. However, traditional Active Contour Models are very sensitive to the position of initial contour,which limits its further application in edge detection area. To overcome this problem, an accurate and fast multi-scale DDCM active contour model is proposed in this paper, where the curvature of the contour is modulated according to different step during edge detection. We employ this approach to detect the edge of interested objects in CT images. Compared with traditional active contour model, the proposed algorithm has a better performance of edge detection with the aspects of accuracy and speed, which is showed by our experiments.2920|187|0更新时间：2024-05-07
- 摘要：This paper introduces a method to pick up inflection points in handwritten signature curves. If wavelet is defined as the second derivative of the Gaussian function, and the original curve is seemed as a continuous function, then the curve got after wavelet transform can be nearly seemed as the second derivative of the original function. It is widely known that where the second derivative of a function equals to zero corresponding to its inflection points, so the inflection points in the original curve can be picked up effectively by detecting zero points in its second derivative curve. Besides, because the wavelet has the advantage that it can provide multiple scales to analysis the signature curves, using this method, handwritten signature can be segmented and matched effectively. The distance between two corresponding segments can be measured with Dynamic Time Warping Algorithm (DTWA), which is a widely used dynamic matching algorithm in speech process and has good performance. Because the segmentation implemented by detecting inflection points has different comparability and stability, combined with DTWA, the method can improve the result of signature verification.2967|201|0更新时间：2024-05-07
- 摘要：Ridge distance is an intrinsic texture property of fingerprints and it has important application in automatic fingerprint identification system. However, up to now, only several literatures touch ridge distance estimation deeply. Ridge distance estimation is developed in this paper and a method based on statistical window for ridge distance estimation in fingerprint images is brought out. A fingerprint image is divided into non-overlap blocks and a statistical window and a base line is defined for each block image. A statistical method is used to acquire the distribution of ridge in the statistical window and the ridge distance of a block image is estimated. Even to block images in which ridge directions vary acutely only if the dominant direction is existent or to block images in which noises are comparatively serious, such as much ridge breaks, the method still can estimate ridge distance reliably. Tests with some typical images show the good performance of the method presented in this paper.2373|165|0更新时间：2024-05-07
- 摘要：Robot Soccer game is become popular recent years. The system combined with mechanic, electronic, control, computer, image process and etc. In order to win the game, the system must be more quick and more accuracy. Because vision sub-system is the first step of the whole system, its speed and accuracy are important to the whole systems speed and accuracy. Real-time segmentation is the first step in the color-vision system on the robot system. A color image segmentation method using improved seed-fill algorithm in YUV color space is introduced in this paper. The new method dramatically reduces the work of calculation, and speeds up the image processing. The result of comparing it with the old method based on RGB color space was showed in the paper. The second step of the vision sub-system is identification the color block that separated by the first step. A improved seed fill algorithm is used in the paper. The implementation on MiroSot Soccer Robot System shows that the new method is fast and accurate.2603|194|0更新时间：2024-05-07
-
The Design and Implementation of Flame Image Processing and Combustion Diagnosis System Based on Web
摘要：In order to carry out the flame visualization and combustion intelligent diagnosis on pulverized coal furnace, and then give the instruction information for power plant staff, we have established a flame image processing and combustion diagnosis system based on Web. In this paper, firstly we propose a kind of browser/server three-layer structure based on Web and Java, then analyse the advantages of this kind of structure in detail. We apply this structure to the flame image processing and combustion diagnosis system . We use optical lens?CCD camera and image card converting the flame video signal to digital signal. By utilizing the flame image processing method and the radiation information of the furnace , we can monitor and analysis the state of combustion and the product of NOX on line. We have developed measurement of project temperature based on colorimetric method and reconstructed section temperature field in terms of ART(Algebraic Reconstruction Techniques). This system has appplied in two of the power plants, effectively improving the level of automatic supervising and safety in the power plants.2368|148|0更新时间：2024-05-07 - 摘要：Aiming at plate texture and plate shape of vehicle image, an algorithm applied to license plate extraction of vehicle image based on mathematical morphology is introduced. It adapts to the complicated background of vehicle image and the variety of illuminating conditions by adjusting threshold, while two-dimensional morphological filtering method. Moreover, point operation of mathematical morphology is inverted into line operation of mathematical morphology, The binarization method is based on one-dimensional edge detection to effectively utilize the plate shape and plate texture feature of vehicle images. It is much more robust and faster than the traditional thresholding methods and edge detecting operator methods. We have completed a series of experiments under different conditions. The experiment results show that the segmentation effects and it's orientation accuracy are improved significantly, so the proposed method is very effective for segmenting and locating noisy vehicle license plates.2902|238|0更新时间：2024-05-07
- 摘要：Looking upon the gray level histogram as a mixture of two Gaussian density functions is a conventional model in the image segmentation, unfortunately the histogram of the complex image often appears a multi-peak feature. In order to get a more accuracy approximation of this kind of histogram, this paper generalizes this model by considering the histogram a mixture of several Gaussian density functions, and employs a new algorithm of Adaptive Oriented Orthogonal Projective Decomposition to handle the mathematical problems involved in this process. In this proposed method, the key parameters of each Gaussian function can be calculated efficiently, which adequately leads to the determination of the optimal thresholds between different neighboring Gaussian functions. A new parameter called the Dividual Ratio of Threshold is introduced and used as the reference for the selection of the final thresholds. Experimental results show that this method can be effectively applied for the multi-threshold segmentation of complex images.2817|213|0更新时间：2024-05-07
- 摘要：To fulfill the demand of some time-critical operations, we presented a distribute algorithm for Boolean operations of sculptured solids. The algorithm exploits parallelism at all stages of the Boolean operation between sculptured solids, and made a good compromise with the relevancy of the topological entities involved in. And different load balancing strategy is used according to different kinds of the distribute tasks undertaken. The algorithm can distribute the notoriously computation cost of Boolean operations between complex sculptured solids on a heterogeneous LAN, and have been implemented in a heterogeneous LAN which is composed by 10 PentiumII/350 PCs and 5 SGI O2 workstations. A good degree of parallel computation can get from the experiment datum. The algorithm is used in a modeling kernel for a feature-based parametric modeling system named GS-CAD, and is able to interactively evaluate the boundaries of Boolean operations of complex sculptured solid in average 3-4 seconds. It can also be easily extended to collision detection algorithm between sculptured solids.2265|208|0更新时间：2024-05-07
- 摘要：Subdivision surfaces generated by subdividing general control meshes with the butterfly subdivision method are theoretically smooth. However, the surfaces are not visually smooth, because they are not fairing. Fairing the subdivision surfaces directly with traditional fairing methods has to store the huge volume data of subdivision results with approximation error introduced. In this paper, a new mesh fairing method is presented to address such problem. The subdivision result is faired through moving the vertices of the initial control mesh other than the vertices of the subdivision surface. With simple constraints, the method can fair the subdivision surface while decreasing the approximation error between the subdivision surface and the real 3D geometrical model. Moreover, this method can reduce the volume of data to be stored since it only needs to store the updated initial control mesh. To generate fairing subdivision surface, one just needs to subdivide the initial control mesh with simple subdivision rules.2688|199|0更新时间：2024-05-07
- 摘要：Cone-beam reconstruction, which is also called volumetric reconstruction, has been one of the central research problems in medical imaging. Compared with traditional 3D reconstruction, it has several advantages, such as faster scanning, better dose usage and the isotropy of the spatial resolution and density resolution in 3D reconstruction object. Based on the idea of Grass's T-FDK algorithm, a new volumetric reconstruction algorithm named FT-FDK, was presented in this paper. The X-ray projection data were collected on a planar detector instead of a cylinder detector, which was quite different from Grass's algorithm. The 2D projects were obtained along a single circular source trajectory. Unlike traditional FDK algorithm, there were three steps in this new algorithm. The first step was that the projection data should be rebinned to oblique parallel projection data. Then, the rebinned oblique parallel projection data should be pre-weighting filtered. And the last step was that filtered projection data should be backprojected to the final 3D reconstructed image data. Theory and experimental results indicated that this algorithm had the same computational complexity as traditional FDK algorithm and the image quality was improved in the reconstruction. The results suggest that this algorithm can be used widely in applications of medical imaging and nondestructive testing.3286|226|0更新时间：2024-05-07
- 摘要：Typically,the image data we get are anisotropic, that is, the distance between adjacent image elements within a slice is different from the spacing between adjacent image elements in two neighboring slices. Interpolation is the key to convert such anisotropic data into isotropic one. The traditional interpolation methods include grey-level interpolation and shape-based interpolation. But both of them have their own shortcomings. Grey-level interpolation is easy to blur the object's boundary and shape-based interpolation is nearly limited to binary images only. In this paper, in order to solve these questions, we present a new way to interpolate grey-level images, which is based on the shape of these images. First, we use mathematical morphology to acquire the contour of the interpolated image. To each point in this contour, we find the corresponding points in both original images. According to the acquired grey value of the two corresponding points, we use linear interpolation to calculate the grey value of the interpolated point. Once we acquire each point's gray value, we obtain the final interpolated image. The experimental results show that the new method is effective.2534|175|0更新时间：2024-05-07
- 摘要：Usually, the construction of 3D Chinese character mainly applies the sweeping principle--that is, the stroke of 3D Chinese character is generated by variational sections sweeping the center line of its 2D stroke. This method exists defects as follow: difficult expression in abundantly variational strokes of Chinese character and huge data in description of the stroke. For this reason, based on the principle of layer-manufacturing in Rapid Prototyping (RP), that is--any 3D Chinese character can be considered as 2D section accumulated layer by layer, this paper put forward a new method to constitute 3D Chinese character by 2D Chinese character. By applying the theory of Voronoi diagram, the offsets of each layer, which are the skeleton of surface of 3D stroke, are automatically generated by basic 2D stroke contours, which is expressed to double circle link table whose node is coordinate vertex of this contour. By making those contours as skeleton and then interpolating block by block in Bezier-B blend spline surfaces, a 3D stroke of Chinese character is constituted and then a 3D Chinese character is composed by this strokes. In the paper, a actual example of 3D Chinese character is displayed at last, and this new method shows that constitution of 3D Chinese character based on layer-manufacturing can more plentifully express the Chinese stroke, and overcome the problem of huge data capacity and distortion when a 3D Chinese character is magnified.2715|165|0更新时间：2024-05-07
- 摘要：Based on the coordinate relationship between the current borderline pixel and its two lockstep borderline pixels, a certain connection by which some pixels on the region connect with the current borderline pixel is defined and constructed here. On the basis of enumerating all the types of the borderline pixels and analyzing their features, an algorithm that ransacks any close region connection by deciding and finding out those pixels in the region corresponding to the current borderline one with that connection, is concluded. Any close region connection means that the close region connection can be either single connectivity one or complexity connectivity one with any borderline pixel curves. That is, the algorithm can deal with almost all close regions in a consistent way. So it has good applicability. Secondly, it gets higher performances in both time and space due to its particular judging rules, retrieving processes and data describing method. Additionally, the operations mainly converge on the integral comparison operation, and the algorithm is simple. In a word, the algorithm meets the region-concerned needs such as region filling, point domain juding, point domain tracking, and so forth. Simultaneously, due to the region pixels describing method, it presents a helpful use for reference for researching how to describing close region more effectively.2566|182|0更新时间：2024-05-07
- 摘要：With the development of digital earth and cyber city, there is an increasing equipment on management and vitalization of integrated 3D model data and image data. How to improve the speed and visualization effect of three-dimesional(3D) model is a major research issue in 3D GIS. In order to improve the speed of 3D visualization, such as flying/walking through, many algorithms have been developed to reduce the data volume of 3D model, for instance, Level -of -Detail(Lod) algorithm. However, image data is also another important factor effecting the operation speed and reality of 3D model. Image data as the texture of surface of 3D model can enhance the reality of 3D model. For the reason that different graphic cards may have different restriction in the size of texture, the texture has to be managed very carefully. This paper proposes a new method to manage image data when a image is used as the texture for a 3D GIS environment. First, the data structure of multi-resolution texture model, which is developed to manage the large texture, is presented. Secondly, the algorithm that creates the multi-resolution texture model is developed, which is of vital importance for the texture mapping. Thirdly, the method of determining a proper texture resolution to map texture on the surface of 3D model based on viewer conditions is discussed. Finally, several experiments are illustrated.3083|181|0更新时间：2024-05-07
- 摘要：Based on the analysis of the color error in topographic map images such as transitional color and color distortion, the deficiency of available algorithms that only make use of color information in map segmentation is indicated. To get correct segmentation, the information in the map must be explored sufficiently. Simulating the integrated process of visual information abstracting, a novel concept of spatial relation information is defined. According to the feature of topographic map the distribution character of pixels relations in the image is analyzed and then proposes the method of how to recover spatial relation information. By the transformation of color spaces and the improved fuzzy clustering algorithm the color of pixels in the map is evaluated with different membership degree. At last spatial relation information and color information is combined in the segmentation of color map. As a result, color error in the map is restrained effectively. The experiments show that the proposed algorithm achieves robust and precise segmentation. Furthermore, the algorithm can recognize some elements in topographic map directly. It has established a fine basis for automatic vectorization of color topographic maps and introduces a new approach for image analysis.2609|161|0更新时间：2024-05-07
- 摘要：Self-calibration is the computation of metric properties of the cameras and the scene from a set of uncalibrated images. This is different from conventional calibration where the camera calibration matrixKis determined from the image of a known calibration grid or properties of the scene, such as vanishing point of orthogonal directions. Instead, in self-calibration the metric properties are determined directly from constraints on the internal or external parameters. Camera calibration is essential to many computer vision applications. In practice this often requires complicated calibration procedures to be carried out regularly. In the last few years a lot of work has been done on self-calibration. It has been shown that a metric calibration was possible based on the rigidity of the scene. Based on the pin-hole camera model,a new self-calibration method is presented in this paper. We can use the char- acteristic of 3-point perspective projective, vanish points and the orthogonal vector to get a set of equations and then all the intrinsic parametersαx,αy,u0,v0can be solved with high accuracy. The feasibility of the approach is illustrated on synthetic images.2664|169|0更新时间：2024-05-07
- 摘要：Camera self-calibration techniques based on active vision make the calibration simplified, so it is a main branch of camera calibration field. Many kinds of distortions were existed in ordinary camera, among these distortions the radial distortion is more serious, so the study on the self-calibration technique taking account of radial distortion is very important. A camera self-calibration algorithm based on active vision taking account of two-degree radial distortion is proposed in this paper in order to make the calibration result more accurate. The epipolar geometry constraint taking account of two-degree radial distortion is developed. Then the conclusion was drawn that the epipole of the image which is taken after the translation motion is done is still equal to the epipole of the initial position image when the two-degree radial distortion is taken account. The calibration taking account of two-degree radial distortion become a problem to solve complex nonlinear equations. The Levenberg-Marquardt algorithm is used to solve the nonlinear equations. Then the intrinsic parameters and two-degree radial distortion coefficients can be calibrated by controlling the camera to undergo four translations or more which should not be co-planar. Experiments results show that the accuracy of the algorithm is high and the robustness of the algorithm is strong. So this algorithm is feasible.2657|156|0更新时间：2024-05-07
- 摘要：In most of the fractal image compression algorithms, the most efficient method to speed up coding process and increase compression ratio is to adopt bigger collage distance. However this method will result in poor image. To overcome the disadvantage, this paper has presented a new fractal compression algorithm based on collage distance fitting and compensating. This algorithm has provided two kinds of different compression codes, which contain their own unique collage distance sets and different types of similar transforms. By applying different types of similar transforms to different range blocks, the collage distance has been compensated, and the image quality and compression speed have been improved. In the end, an experiment of 'Lena' image has provided the application and demonstration of this algorithm. Theoretical analysis and application in practice have indicated that the proposed algorithm can be regarded as a suitable alternative to ordinary fractal compression method for balancing between the distortion measure and compression ratio.2697|369|0更新时间：2024-05-07
- 摘要：In this paper, a new approach for optimization design of biorthogonal wavelet filter in image compression is presented. In the process of image encoding, we start with the principle of choosing wavelet bases , make the widely used Antonini's 9/7 wavelet basis the baseline solution, exploit the image compression method of the set partitioning in hierarchical tree(SPIHT), find the optimization filter coefficients of a biortherogonal filter bank that gives the image compression the best performance. We use Lena as the standard train image, make the compression rate as parameter, apply the genetic method and find the optimization of biorthogonal wavelet filter banks that maximizes the PSNR. It is an optimization problem. In train and test, the results are that the image compression performance is better to use the new optimization biorthogonal filter banks than the Antonini's 9/7 biorthogonal filter banks.2794|163|0更新时间：2024-05-07
- Postal code：100190
- Tel：010-58887035/58887030/58887418 Email：jig@aircas.ac.cn
- Technical support is provided by Beijing Founder electronics co., LTD 京ICP备05080539号-4
京公网安备11010802024621 - It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
- Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.
0