最新刊期
卷 17 ， 期 3 ， 2012
- 摘要：Collinearity equations play a very important role in photogrammetry and computer vision. It established the relationship between three points, that is, the camera projective center, an object point and its corresponding image point. In collinearity equations, a rotation matrix is usually used to describe the attitude of an image traditionally a rotation matrix is always represented by three Euler angles because of its vivid describing of the relationships between three axes. However, since the unit quarternion was presented by Hamilton in 1843, its use has extended into many application fields such as signal processing, mechanics, and aerospace. In order to discuss the application problem of unit quaternion in photogrammetry, research was carried out in this paper systemically into rigorous solution of collinearity equations by using unit quarternion based rotation matrices. Starting from the basic theories and its operations of unit quaternion, rigorous linearized expression of unit quaternion based collinearity equations are derived in detail and there is no need to derive the rotation matrix. Tests by using both simulated data and real image data indicate that linearized collinearity equations have many merits such as having a very simple form, being independent of initial values, and having a high convergent speed compared to two other rotation matrices orthe Euler angle based rotation matrix. So, unit quaternion based rotation matrix should be used widely in practical application.关键词：collinearity equation;unit quaternion;rotation matrix;bundle adjustment3596|215|0更新时间：2024-05-08
- 摘要：In the research field of digital signal processing,compressive sensing (CS) becomes more and more important because it changes the traditional signal processing method based on Shannon’s sampling theorem.Under the CS framework,signal recovery is a key point to obtain the digital termination product.The basis pursuit (BP) algorithm seems the most fundamental method of CS recovery,which is essentially an L-norm minimization problem.However,BP can not be used for the signals with more than one dimension.Therefore,this paper presents a new high-dimension CS recovery method based on fractal dimension theory.The Minkowski dimension is used to replace the L-norm as an object function in CS recovery.The visualization and SNR of our experimental results show that fractal CS recovery not only inherits the advantage of BP but also improves the dimensional extensive property.关键词：compressive sensing;signal recovery;L;-norm minimization;fractal dimension3091|186|0更新时间：2024-05-08
- 摘要：In the encoding process of vector quantization (VQ), a great deal of distance computations between vectors are needed, which is computationally expensive and prevents its practical applications. In order to speed up the VQ encoding process, lots of fast codeword search algorithms based on 1-D characteristics have been proposed for reducing the codeword search space. This paper aims at improving the method using inequality proposed by Mu et al by using more effective method based on 1-D characteristics and adaptive subvector povrtitions. The experimental results show that our proposed scheme can reduce the extra memory requirment from (-1)/2 to 13, meanwhile reducing the codeword search space by 33.88%50.94% and reducing the encoding time by 10.82%27.16%.2712|122|0更新时间：2024-05-08
- 摘要：We propose an improved anisotropic diffusion method based on the human visual model to gradient threshold for the diffusion coefficients.First,we studied the evaluation of gradient thresholds for anisotropic diffusion coefficient based on the primal PM equation.Through analyzing Weber’s law according to the bionics principle and based on the background,the image brightness was divided into different regions.The gradient thresholds for the different areas were calculated according to the different equations.Then,we discussed the discrete implementation of the proposed method for the eight neighborhoods.The simulation results indicate that the proposed method not only preserves the image information better,but also suppression the noise more satisfactory.关键词：anisotropic diffusion;diffusion coefficients;gradient threshold;bionics principle;Weber law3459|187|0更新时间：2024-05-08
- 摘要：The polyphase matrix decomposition coefficient of Integer Lifting Wavelet is is not unique.It has diverse selection methods and large computation quantity.First,the selection algorithm of filter iteration times is adopted to obtain the optimal iteration times,according to the input signal-to-noise ratio.Then,the comparison algorithm of the nonlinear iterated function,which combines the optimal iteration time,is treated as criteria to obtain the optimal decomposition coefficient.The iteration time is based on the input data,so the decomposition coefficient will have the optimal processing effect for the data,meeting the selection requirements of the polyphase matrix decomposition coefficients.The comparison algorithm meets the convergence characteristics,which can reduce the number of iterations by comparing whether the impulse response and step response of filters meet the error limit,in order to obtain the optimal decomposition coefficient quickly.Stemming from experiment results,this fast extraction algorithm can effectively meet the requirements of data processing,reduce the data computation quantity and elevate the data processing efficiency.关键词：integer wavelet transform;iteration times;polyphase matrix;fast extraction algorithm3187|150|0更新时间：2024-05-08
- 摘要：In the traditional examplar-based algorithm,the image completion effect is greatly degraded due to the fixed patch-size,which cannot change together with an image automatically in the process of completing.So in this paper we propose an adaptive patch-size determining method to solve this problem.By analyzing the image changes in the gradient domain,this method acquires the structure information of each pixel,and then automatically adjusts the patch-size.The simulation results show that this method can overcome many shortcomings of traditional methods,such as structural error propagation and whole structure loss,and realizes perfect completion effect for images with obvious changed structure.关键词：image completion;texture synthesis;adaptive patch size;gradient field3275|224|0更新时间：2024-05-08
- 摘要：We propose an improved(FCM)approach for interactive image segmentation to solve the problem that the objective contour is easily influenced by natural texture.Users interactively input seed points as clusters centers to get the main feature differences between foreground and background;a global spatial similarity measure model and Gabor energy filters are introduced into the distance metric to measure the similarity between all the pixels and the cluster centers.Edge density is introduced into the definition of the weight factor,according to image features,the ratio of the texture feature and the color feature in the feature space for pixels which can be adaptively calculated,making the feature we calculated more accurately reflect the essential attributes of the image.In this paper,experiments have been carried out on images with natural texture background,and two performance measures were used to compare the proposed algorithm with the random walker(RW)algorithm.It is shown that the accuracy of the proposed method surpasses FCM and RW methods.关键词：fuzzy c-means(FCM);image segmentation;gabor energy filters;edge density;random walker3198|234|0更新时间：2024-05-08
- 摘要：In order to identify transmission lines accurately from low-contrast images with complex backgrounds,we applied the total variation denoising model.A two-dimensional explain based on the Simulated Annealing and Particle Swarm Optimization Algorithm is proposed and used for image segmenation.Besides,the traditional Freeman Chain Code Representation Method is improved to extract the transmission line target,and a line fitting method based on the Least Square Method is used to recover missing parts of the basic skeleton of the transmission lines.The experimental results indicate that the two-dimensional OTSU based on the Simulated Annealing and Particle Swarm Optimization Algorithm has better segmentation effect with better convergence and higher computing speed in the search of the best segmentation threshold.The transmission lines extraction algorithm based on the improved traditional Freeman Chain Code Representation Method can remove the complex image background well and extract transmission lines completely.关键词：transmission lines;image denoising;image segmentation;line fitting2831|148|0更新时间：2024-05-08
- 摘要：The detection and tracking of small and/or indistinct objects is a challenge in computer vision, since it is difficult for current detection and tracking algorithms to extract appropriate apparent features from small and/or indistinct objects. We found by observing video data that the small moving objects can be distinguished from noisy background due to their regular motion pattern. Based on this finding, we proposed a novel method for solving the challenging problem of small and/or indistinct object detection and tracking. The contribution of this paper is the denoising algorithm for small object segmentation and the dynamic programming algorithm for route optimization in tracking. Extensive experiments show the promising results in both accuracy and robustness of the proposed method.关键词：foreground;association;denoising;dynamic programming3803|196|0更新时间：2024-05-08
- 摘要：A new smalltarget detection method under sea surface environment is presented.Because the frequency tuned saliency detection is based on differences of the mean values of image features and the result of Gaussian filtering,the performance is not satisfying when there are many clutters in the background.The proposed approach makes improvement to frequency tuned saliency detection.The three components of image in LAB color space are divided into patches and within each patch,the frequency tuned saliency detection is conducted.The master saliency map is the fusion of each patch’s saliency map.The proposed method overcomes the limitation of frequency tuned saliency detection when there are many clutters under sea background.The effectiveness of it is demonstrated by experiments.关键词：visual attention;frequency tuned;saliency;small target detection under sea surface environment3170|157|0更新时间：2024-05-08
- 摘要：Pedestrian detection in images with complex backgrounds is valuable in theory and applications.To deal with the variability of illumination and pedestrian’s poses,the local feature based on vision attention (LFVA),derived from a visual saliency mechanism,is proposed in this paper.LFVA is illumination and rotation invariant,and can also be used for multi-scale analysis.A pedestrian model based on feature blocks is proposed,in which the pedestrians are represented by a set of feature blocks.Every feature block is represented by the position and histogram of the LFVA.The pedestrian model is obtained by clustering.The AdBoost detection classifier is trained using the maximal responses of the feature blocks and is improved using hard negative samples and trusted samples.The local maximal of the response of the detection classifier in image and scale space is located as the position of the pedestrian by sliding window search.Compared with the existed methods,the proposed method is less sensitive to vertical edges and can deal with occlusion and pose variation to some extent.关键词：vision attention;local feature;pedestrian detection;feature block3395|266|0更新时间：2024-05-08
- 摘要：Face recognition has been always a hot spot in image processing and pattern recognition. However,problems,such as small samples,rotation,and scale changes,always restrict its development. In this paper,a scale invariant feature and template based method is proposed to achieve recognition upon a single face sample. First,scale invariant features are extracted and matched to obtain feature-level correspondence between faces,upon which rotation and scale variances between faces can be estimated and rectified for invariant templates. Then a sub-region based zero mean Nermaczed Cross-Correlation (ZNCC) method will be adopted to acquire region-level correspondences between the templates and rectified faces. Finally,a joint judgment will be made according to both feature-level and region-level correspondences between faces. This method is feasible and robust not only to rotation,scale and illumination variance,but also to low texture.关键词：scale invariant feature;rectification;template;ZNCC;face recognition3374|217|0更新时间：2024-05-08
- 摘要：For intelligent transportation systems, vehicle recognition in complex traffic scenes is a key issue. In this article, a novel scheme using(HOG) features and sparse representation target recognition for vehicle recognition in complex traffic scenes is proposed. Our method uses the HOG to extract features from samples and candidate targets, and then wses trained samples as an overcomplete dictionary based on sparse representation. Finally, candidate targets are recognized by computing sparsity and reconstruction residuals in the dictionary. Experiment results show that the proposed scheme provides higher recognition preciseness in real time, even in complex traffic scenes such containing occlusion and a large variety of target classes.关键词：sparse representation;histograms of oriented gradient(HOG);vehicle recognition;intelligent transportation;compressive sensing3012|239|0更新时间：2024-05-08
- 摘要：Images are the main source through which humans acquire and exchange information. Edges are one of the most fundamental and important feature of image. Edge detection is a classic research problem of computer vision and image processing. In this paper we present a new edge detection operator based on a composite derivative,which is realized by the combination of fractional differentiation and integration. The features of the new operator,in terms of detection accuracy and noise immunity,are demonstrated through experiments,and the experimental results verify the effectiveness of the new operator. Finally,qualitative and quantitative comparisons of the new operator with two classical operators,Canny and CRONE,are performed. The comparisons show the superiority of the new algorithm,which is reflected in the abilities of accurate edge localization and good suppression of false edges while still being able to detect fine true edges.关键词：image processing;edge detection;fractional differentiation and integration;detection accuracy;noise immunity;qualitative and quantitative comparisons.2985|201|0更新时间：2024-05-08
- 摘要：To improve detection accuracy and speed,a corner detection method based on the novel Local Corner Response Saliency (LCRS) and Iterative Segment Approaching (ISA) is proposed.First,a new metric,the Local Corner Response Saliency,is defined to measure the response saliency of a candidate corner point within the local region where it is located.The corner detection criterion based on LCRS proved to be equivalent to a local adaptive thresholding method.Then,by treating LCRS as a property of response saliency of a region,the task of searching corner points is transformed to the task of searching special regions with high saliency.With this point of view,an iterative segment strategy is suggested to gradually narrow down these salient regions and finally approaching the actual positions of the true corners.The ISA algorithm can work with various definition of corner response function (CRF),and its speed in average case is identical to the Harris detector.Experiments show that when using Noble’s CRF,the average false and miss detection rates of ISA are lower than those of Noble detector by 4.62% and 5.59%,respectively.When using the Harris’ CRF,these tow rates are also lower than those of the Harris detector by 2.87% and 3.37%,respectively.Moreover,in both cases the average running time of ISA is shorter than these of the Harris and Noble detectors.关键词：corner detection;local corner response saliency;iterative segment approaching;Harris detector;Noble detector2849|185|0更新时间：2024-05-08
- 摘要：Accuracy is important for the regrstration of medical images. Pixel gray values are a widely used feature in image registration. However,the gray values come from a single source and ignore the spatial information. In some cases,it will cause misalignment. To solve the problem,entropic graph estimation integrated with SIFT features is proposed as a medical image non-rigid registration algorithm. In the algorithm,mutual information based rigid registration is used to roughly register two images. Then the pixel gray value and the SIFT features are extracted to form a k-nearest neighbor graph (kNNG),which is used to estimate α-mutual information (αMI). Comparison results of the images obtained from lung CT images and brain MRI images showed that the proposed algorithm provides better accuracy than both,the conventional rigid registration algorithm based on mutual information and the non-rigid registration algorithm based on entropic graph estimation and single pixel gray values.关键词：medical image registration;SIFT (scale invariant feature transform) descriptor;k-nearest neighbor graph (kNNG);α-mutual information (αMI)2786|196|0更新时间：2024-05-08
- 摘要：An improved 3D surface reconstruction method is proposed for non-uniform sampling points. The method performs spatial partitioning for an entire set of points,in order to reduce the search range of neighbor points and decrease the search time. For searching topological neighbor points,geometric neighbor points are calculated,and Minimum Spanning Trees are constructed by finding directional points. After projecting topological neighbor points onto local tangent planes,constrained triangulation is carried out for the projected points. Then the connection of projected points is mapped directly back onto 3D space. As a result,the 3D surface is reconstructed successfully. The experimental results show that the improved algorithm is efficient,has good reconstruction effects,and can widely be used for surface reconstruction of non-uniform sampling points.关键词：non-uniform sampling points;surface reconstruction;directional point;topological neighbor point;triangulation2974|187|0更新时间：2024-05-08
- 摘要：To satisfy the demands for 3-D reviewing in plant design,a multi-resolution rendering approach for large-scale process plant models based on programmable graphics pipelines is proposed.Based on the analysis of the characteristics of a process plant model,a multi-resolution model of the object is constructed according to the multi-resolution templates of the basic voxels.In the generation process of the multi-resolution, model the proposed approach can ensure that the vertex set of the low-resolution model is the subset of the vertex set of the high-resolution model.There fore,only the vertex information of the highest resolution model and some of the vertex indexes need to be saved for every object.On this basis,with the effective management of the vertex buffer and the index buffer,a multi-resolution model transform can be implemented according to the vertex index transform in the rendering process.In addition,the data exchange between main memory and display memory can be reduced.Examples show that the approach presented in this paper can be implemented on a general PC and satisfy actual work demands from the designer.It can preprocess a 21 M triangles model within 10 and achieve 31 frame/s smooth frame rates on average,without using extra hard-disk space.关键词：multi-resolution model;large-scale and complex scene;process plant;programmable graphics pipeline;3D review2559|136|0更新时间：2024-05-08
- 摘要：The reflection education of the environment mapping technique is geometrically biased and cannot obtain parallax effects because the object’s surface position is unknown. We improved the technique by introducing position rectification. First, the environment’s position information is stored in a distance map; then the query vector of the environment map is rectified in two conditions using the distance map to reduce the geometrical bias in the calculation of the reflection.A layered rendering method is given to manipulate scenes that contain multiple objects.Our results show the introduction of the position rectification boosts the accuracy of environment mapping in reflection calculation and can obtain the parallax effect caused by the occlusions of objects.关键词：environment map;distance map;parallax2818|137|0更新时间：2024-05-08
- 摘要：Extraction shoreline information from remotely sensed imagery is important for islands investigation With remote sensing and the coastal comprehensive management. In order to accurately extract the information of shorelines,a new algorithm was developed based on a "bottom-up" approach cellular automata (CA) by considering the direction information of images. The algorithm was developed under a Matlab environment,which includes functions such as mean smooth,land/water segmentation,discrete objects removal,and shoreline tracing.Using the proposed algorithm,the shorelines of an artificially coast at the Changxing-Hengsha islands of Shanghai and a muddy coast at Chongming Dongtan of Shanghai were extracted from Landsat ETM+ images.The results demonstrate that the CA based shoreline extractor can accurately detect shoreline information of different coast types (e.g.artificially coast and muddy coast) from remote sensing images.关键词：shoreline extraction;cellular automata;direction information;remote sensing imagery3566|189|0更新时间：2024-05-08
- 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