202 results about "Kernel sparse representation" patented technology

The invention discloses an image super-resolution reconstruction method based on sparse representation and deep learning, and solves the problems that the image super-resolution process is complex incalculation and the quality of a reconstructed image is poor. The method comprises the following implementation steps: collecting and extracting training data blocks and a chromaticity and brightnessdictionary for combined optimization training; independently reconstructing a high-resolution image block; Carrying out high-resolution image reconstruction of sparse representation; Training a residual error network based on deep learning to optimize high-frequency details; Image super-resolution reconstruction. In order to prevent an edge effect and a fuzzy effect, chroma and brightness data aredistinguished and independently reconstructed; in order to optimize high-frequency detail information of a sparse representation output high-resolution image, the high-resolution image based on sparse representation reconstruction is input into a residual network, and a high-frequency residual image is output through four times of convolution feature extraction and feature fusion and input bitwise addition to reconstruct a super-resolution image. The method is low in calculation complexity, high in image reconstruction quality and widely applied to the fields of remote sensing monitoring, criminal investigation, traffic management and the like.

The invention discloses an SAR image target recognition method based on sparse representation. The SAR image target recognition method based on sparse representation mainly resolves the problem that an existing method is complex in preprocessing and difficult in estimation of an azimuth angle. The SAR image target recognition method based on sparse representation comprises the steps of (1) extracting partial features of an image and studying a recognizable dictionary through a diversity density function; (2) carrying out sparse encoding on each partial feature through the dictionary, and then carrying out space pooling on each divided sub-area through a space domain pyramid structure to obtain feature vectors of the sub-areas samples of a training set and a test set; (3) weighing the corresponding sub-areas of a test sample according to the sparsity of each sub-area of the test sample; and (4) combining the weighed sub-areas together and recognizing the image through a sparse representation method. Compared with the prior art, the SAR image target recognition method based on sparse representation has high robustness for shielded and partial noise, improves the recognition accuracy of an SAR target without estimating the azimuth angle, and can be used for image processing.

The invention discloses a face identification method based on multiscale weber local descriptor and kernel group sparse representation. The face identification method comprises the following steps: firstly normalizing the size of face images and smoothing the images by utilizing a gaussian filter; extracting differential excitation ingredients of the multiscale weber local descriptor of the images and extracting direction information by utilizing an Sobel operator; extracting the multiscale weber local descriptor of the face images according to the multiscale differential excitation and the direction information and mapping the multiscale weber local descriptor to a kernel space by utilizing a histogram intersection kernel; then with a kernel matrix obtained by a training sample as a sparse dictionary, calculating group sparse representation coefficients of a kernel vector obtained by a test sample; and finally reconstructing a multiscale weber local descriptor vector of the test sample according to the group sparse representation coefficients and distinguishing the test sample by utilizing the minimum reconstruction error. According to the face identification method, the multiscale weber local descriptor and the kernel group sparse representation algorithm are fused for face identification, and the identification accuracy rate is greatly improved.

The invention discloses a sparse representation and dictionary learning-based bearing fault classification method. The method is characterized by comprising the following steps of: preprocessing acquired history bearing vibration signals to obtain vibration signals under different working states, respectively constructing corresponding sub-dictionaries and combining the sub-dictionaries into a redundant dictionary; and online acquiring a bearing vibration signal by using a sensor, solving a sparse coefficient of the signal under the redundant dictionary by using a generalized orthogonal matching pursuit algorithm, and classifying the vibration signal through a reconstruction error so as to identify the bearing working state. According to the method, relatively good classification effect can be obtained, the dictionary training process can be accelerated and the ability, of being adapted to target signals, of the dictionary is improved, so that the method is capable of realizing the sparse decomposition of complex vibration signals more efficiently and is used for fault identification.

The invention discloses a hyperspectral abnormal object detection method based on structure sparse representation and internal cluster filtering, aiming at addressing the technical problem of low object detection effciency of current hyperspectral abnormal object detection methods. The technical solution involves: after selecting an initial background pixel, using the dictionary learning method which is based on principal component analysis to study a background dictionary which obtains rebustness, in the course of sparse vector resolution and image reconstruction, introducing re-weighted laplacian prior to increase the solution precision of sparse vector, computing the errors betwen an original image and a reconstructed image to obtain a sparse representation error, using the internal cluster filtering to represent space spectrum characteristics of hyperspectral data, obtaining the internal cluster error by computing the error between a to-be-tested pixel and other pixel linear representation result, and finally combining the sparse representation error and the linear weighting of the internal cluster error and implementing precise extraction of an abnormal object. According to the invention, the method increases 10-15% of detection rate with the proviso of a constant false alarm rate compared with prior art.

The invention discloses a super-pixel polarimetric SAR land feature classification method based on sparse representation. The method comprises: inputting polarimetric SAR image data to be classified, processing the image, and thereby obtaining a pseudocolor image corresponding to Pauli decomposition; performing super-pixel image over-segmentation on the pseudocolor image to obtain a plurality of super-pixels; extracting features, which are seven-dimensional, of radiation mechanism of the original polarimetric SAR image as features of every pixel; performing super-pixel united sparse representation to obtain sparse representation of each super-pixel feature; classifying by using a sparse representation classifier; working out the mean value of each super-pixel covariance matrix, then performing super-pixel complex Wishart iteration by using the classifying result in the last step, and at last obtaining a final classifying result. According to the super-pixel polarimetric SAR land feature classification method based on sparse representation, the problem that traditional classifying areas based on the single pixel are poor in consistency is solved, and operating speed of the algorithm is greatly increased on basis of improving accelerate.

The invention relates to an image super resolution rebuilding method based on sparse representation and various residual. The image super resolution rebuilding method based on sparse representation and various residual includes the steps: first, calculating residual between images formed after the existing high resolution image and a low resolution image are amplified through an interpolation and calculating a high frequency portion and a low frequency portion of the residual; second, building a sample pair through low resolution image sample characteristic and corresponding the high frequency portion and low frequency portion of the image residual, utilizing a texture meta structure to classify the samples by regarding the low resolution sample as a standard and utilizing a singular value decomposition (KSVD) method to train each type of a sample to obtain a dictionary pair of the low resolution sample, the high frequency portion and low frequency portion of the image residual; finally, choosing the dictionary pair and combing the final image residual with the interpolation result of the low resolution image to obtain a high resolution image according to the texture meta structure type of the test samples. The image super resolution rebuilding method based on sparse representation and various residual just needs to rebuild the image residual, combines the interpolation image to rebuild the high resolution image and improves a rebuilding result of the high resolution image.

The invention provides an infrared small-target detection method based on mixing Gauss and sparse representation. The method comprises the following steps: constructing an overcomplete morphological dictionary of an image in a self-adaptive mode by adopting a K cluster singular value decomposition K_SVD method; according to the characteristic that object signals are usually distributed in a Gauss mode, dividing the atoms of the self-adaptive overcomplete morphological dictionary into target atoms representing target forms and background atoms representing background noise components by using a Gauss overcomplete dictionary, and forming a self-adaptive mixing Gauss overcomplete dictionary having a target morphological dictionary and a background morphological dictionary; performing sparse representation on an original image block in the mixing Gauss overcomplete dictionary and extracting the sparse representation coefficient of an image signal; and when the rarefication degree represented by the sparse representation coefficient is greater than a threshold, determining that the image block contains a target, otherwise determining that the image block is a background. By using the method provided by the invention, defects can be overcome that it is difficult for a conventional Gauss sparse dictionary to be adaptive to non-Gaussian distributed target forms and judging whether a target is contained by Gauss atom sparse representation coefficients, and thus the detection performance of small and weak targets can be improved.

The invention provides a wavelet transform and joint sparse representation-based infrared and invisible light image fusion method, and relates to the field of image fusion. The method comprises the following steps of: firstly, carrying out DWT transform on a source image, decomposing the source image into a low-frequency sub-band matrix coefficient and a high-frequency sub-band coefficient, decomposing the low-frequency sub-band coefficient into a matrix by using a sliding window strategy and learning a dictionary in allusion to the decomposed low-frequency sub-band matrix; secondly, respectively fusing the low-frequency sub-band coefficient and the high-frequency sub-band coefficient; and finally reconstructing a fused image through DWT inverse transform. According to the method, effective sparse representation can be carried out on outstanding detail features of source images, and multi-scale fusion can be carried out on detail information of the images, so that target information of infrared images and background information such as details, profiles and the like of invisible light images are well retained, the target identification ability is improved, and benefit is brought to subsequent processing systems to extract and use the information; and compared with the traditional wavelet transform-based fusion method and the existing joint sparse representation-based fusion method, method provided by the invention has advantages.

The invention discloses a terror video identification method and device based on sparse representation of a context. The method includes the steps that first, shot segmentation is performed on each training video sample, and then a pair of key frames are selected according to each shot to represent the shot; visual characteristics of each key frame are extracted, and audio characteristics of the whole training video samples are extracted; the graph of the contact relations between all the key frames in each training video sample is established; the visual characteristics and the audio characteristics of the video to be identified are extracted; a price matrix between the video to be identified and the training video samples is constructed; based on a sparse representation model of the context, the video to be identified is reconstructed through all the training video samples, and the type of the training video samples smallest in reconstruction error is the type of the video to be recognized; according to a sparse representation module of the context, the context relation graph of the training video samples serves as a sparse representation dictionary and is restrained through the price matrix.

The invention discloses a structure sparse representation-based remote sensing image fusion method. An adaptive weight coefficient calculation model is used for solving a luminance component of a multi-spectral image, similar image blocks are combined into a structure group, a structure group sparse model is used for solving structure group dictionaries and group sparse coefficients for the luminance component and a panchromatic image, an absolute value maximum rule is applied to partial replacement of the sparse coefficients of the panchromatic image, new sparse coefficients are generated, the group dictionary and the new sparse coefficients of the panchromatic image are used for reconstructing a high-spatial resolution luminance image, and finally, a universal component replacement model is used for fusion to acquire a high-resolution multi-spectral image. The method of the invention introduces the structure group sparse representation in the remote sensing image fusion method, overcomes the limitation that the typical sparse representation fusion method only considers a single image block, and compared with the typical sparse representation method, the method of the invention has excellent spectral preservation and spatial resolution improvement performance, and greatly shortens the dictionary training time during the remote sensing image fusion process.

The invention discloses a non-reference image quality evaluation method based on discrete cosine transform and sparse representation to mainly solve the problem that in the prior art, non-reference image quality evaluation is not accurate. The method comprises the following steps that a gray level image is input, discrete cosine transform is carried out on the gray level image, and natural scene statistical characteristics are extracted; natural scene statistical characteristics of a series of images of different distortion types and different content are extracted, and an original characteristic dictionary is established according to the average subjective difference score; clustering is carried out on the original characteristics dictionary, and atoms are selected in a self-adaptation mode according to the tested image characteristics and the approximation degrees in the original characteristic dictionary to form a sparse representation dictionary; the tested image characteristics are solved and the sparse representation coefficients are calculated through sparse representation in characteristics space, linear weighting summation is carried out according to the subjective evaluation values in the sparse representation dictionary, and the image quality measure is obtained. The method has good consistency with the subjective evaluation result and is suitable for quality evaluation on images with various distortion types.

The invention discloses a vision target tracking method based on sparse representation and belongs to the technical field of computer vision. The method includes the steps that firstly, a judgment dictionary, a matching dictionary and a gray level matrix T which are needed for following tracking processing are determined based on target images, image frames to be tracked are tracked, multiple first candidate image sample sets are sampled, K representative cluster centers are selected through K mean value clustering, the confidence values of the cluster centers are calculated based on the judgment dictionary, a sample center is set based on the largest confidence value so as to acquire second candidate image sample sets, N candidate images with the highest confidence values are selected from the sets, a gray level matrix of the candidate images is sampled on the basis of the matching dictionary and through fragments, and the candidate image with the highest similarity to the gray level matrix T serves as the tracking target of the current frame. The method is applied to the intelligent monitoring field and has the high robustness on the aspects of tracked target posture change, environment illumination change and blockage and the like.

The invention discloses a super-resolution image reconstruction method based on structure self-similarity and sparse representation. The method includes the main steps of firstly, filtering a set of training sample image to extract features; then, extracting small patches to construct a dictionary including a high-resolution image block and a low-resolution image block in pair, conducting interpolation amplifying on an inputted low-resolution image, conducting filtering to extract the features, solving a reconstructed weight matrix W, conducting iteration to renew a sparse coefficient {alpha i} and a high-resolution image X to be reconstructed; finally, recovering a satisfying high-resolution image till the iteration is convergent. According to the method, the structure self-similarity of the image is used for solving the problem that an existing method is not high in quality. The operation time is short, the efficiency of image reconstruction is high, the quality of the reconstructed image is high, and various natural images which include non-texture images such as animal and plant images and human images and strong-texture images such as architecture images can be reconstructed.

The invention relates to a sparse representation-based image super-resolution reconstruction method. The steps include: 1) selecting a part of an input image sequence as a salient region, the remainder being a non-salient region; 2) training a pair of salient dictionaries D'1 and D'h according to the salient area, and performing context sparse decomposition on the salient area to obtain a salient sparse coefficient on the low-resolution salient dictionary; 3) training a pair of general dictionaries D1 and Dh according to the non-salient area, and performing sparse decomposition through the low-resolution general dictionary D1 to obtain a non-salient sparse coefficient; and 4) multiplying the sparse coefficient by the high-resolution salient dictionary D'h or the high-resolution general dictionary Dh to perform ratio reconstruction, thereby obtaining a high-resolution image sequence. On the basis of a traditional sparse representation super-resolution frame, the sparse representation-based image super-resolution reconstruction method emphasizes internal structure information of an image, and uses the internal structure information of the image as a prior model constraint L0-norm problem to solve, and the performance is superior to other methods in subjective and objective effects while complexity equivalent to that of a traditional sparse representation method is maintained.

The invention discloses and particularly relates to a remote sensing image cloud removing method based on sparse representation. The remote sensing image cloud removing method comprises the following steps of: extracting a cloud mask matrix to acquire a new image; initializing iteration parameters; fixing a texture part of the image and updating a smooth part of the image; fixing the smooth part of the image and updating the texture part of the image; adjusting the smooth part by virtue of a total variation function; updating a iteration threshold; and judging whether decomposition is completed. The remote sensing image cloud removing method based on the sparse representation can adjust the parameters of the total variation adjusting function of the image, namely, a convergence rate of an algorithm is adjusted by changing the parameters of the total variation adjusting function when the block coordinate relaxation algorithm is adopted foe decomposing the image, so that the efficiency of the algorithm and the decomposing effect of the image are improved. The remote sensing image cloud removing method has an obvious thick cloud removing effect on the premise of not destroying original information of the image as much as possible.

The invention provides a color image authentication method and system based on hypercomplex number encrypted domain sparse representation. The color image authentication method comprises the steps of generating a phase mask according to a Logistic chaotic sequence, carrying out double random phase encryption on a color image by adopting quaternion Fourier transform so as to acquire ciphertext data, carrying out sparse representation on a component of the ciphertext data by using a sparse matrix so as to acquire a sparse ciphertext, wherein the component comprises a real component and an imaginary component, decrypting the sparse ciphertext, calculating a nonlinear correlation coefficient of a decrypted image, and carrying out authentication on the color image according to a centralized peak value. The color image authentication method based on the hypercomplex number encrypted domain sparse representation in the embodiment of the invention has high security, and a certain degree of noises can be resisted, thereby being suitable for being applied to the fields of secret communication and authentication of images.

The invention discloses an image sparse representation method based on a Curvelet redundant dictionary, mainly aiming to solve the problems that in the existing method, the redundant dictionary has large scale, the calculation complexity is high, and sparse representation can not be effectively carried out on the rich border outline details in the image. The invention is realized through the following steps: (1) selecting the tight frame of Curvelet as an atomic model; (2) determining the numeric areas of the scale parameter j, direction parameter theta and displacement parameter k in the frame, carrying out discretization on each parameter to form the Curvelet redundant dictionary; and (3) blocking each input image, carrying out sparse decomposition on each sub-image by utilizing an orthogonal matching pursuit (OMP) algorithm sparse decomposition to solve sparse coefficient vectors, combining all the sparse coefficient vectors to obtain the sparse matrix, and multiplying the sparse matrix by the Curvelet redundant dictionary to obtain the sparse representation results of the input image. Compared with the prior art, the invention has the advantages of low calculation complexity, high quality of sparse representation image, especially can better capture the singularity of curves in the image, and can be applied to the fields of image processing and computer vision.

The invention discloses a sparse representation and empty spectrum Laplace figure based hyperspectral data dimension reduction method which mainly aims at solving the problems that the traditional manifold learning information is single and large-scale data are difficult to be processed. The sparse representation and empty spectrum Laplace figure based hyperspectral data dimension reduction method comprises the steps of step 1, selecting a certain amount of data from large-scale hyperspectral data to serve as a training sample; step 2, performing construction of an empty spectrum Laplace figure on the training sample; step 3, performing characteristic decomposition on a Laplacian matrix to obtain the low-dimension representation of the training sample; step 4, constructing a high-dimensional dictionary and a low-dimensional dictionary through the training sample and the low-dimension representation of the training sample; step 5, calculating sparse representation coefficients of remaining hyperspectral data on the high-dimensional dictionary; step 6, performing multiplication on the sparse representation coefficients and the low-dimensional dictionary to obtain the low-dimension representation of the remaining data; step 7, integrating the low-dimension representation of the training sample and the remaining data to obtain complete dimension reduction data. According to the sparse representation and empty spectrum Laplace figure based hyperspectral data dimension reduction method, the effect of the manifold dimension reduction is improved and the large-scale hyperspectral data can be processed.

The invention discloses a sparse representation-based single lens calculation imaging PSF estimation method. In view of requirements on the PSF estimation speed and the precision in single lens calculation imaging, a sparse representation method is tried to estimate the PSF. Firstly, a clear image in a target function is represented to be a product of an over-complete dictionary and a sparse coefficient, and the sparse coefficient is constrained; then, a fuzzy kernel, the over-complete dictionary and the sparse coefficient are sequentially and alternatively estimated through iterative optimization algorithm, and thus the fuzzy kernel needed for single lens calculation imaging can be obtained. In the case of training the over-complete dictionary, a fuzzy image obtained by the single lens imaging system only needs to be known, more extra information is not needed, and as multiple times of convolution operation can be avoided during the calculation process, influences on image edge information restoration are reduced. The method is simple to operate, and is of great significance in the fields of image processing and camera design.

The invention discloses a no-reference color image quality assessment method on the basis of the sparse representation and belongs to the technical field of digital image processing. For images to be evaluated, the no-reference color image quality assessment method includes selecting an image block with the maximum pixel spatial correlation difference from an HSV (hue, saturation and value) color space, and matching elements in a dictionary by the image blocks, wherein each element in the dictionary is composed of the image block with the maximum pixel spatial correlation difference in a training image set, DMOS (differential mean opinion score) and the maximum difference; adopting a matching and tracking method to obtain sparse representation of the images to be evaluated in the dictionary, and setting up the color image quality assessment indexes by the DMOS and the maximum difference of the selected elements from the dictionary. The element dictionary forming method is simple and clear, the extracted characteristics more conform to visual perception of human eyes, the integral algorithm is simpler, and the quality assessment results are more accurate.

The invention discloses a tracking method on the basis of multitask combined sparse representation, which comprises the following steps of: when starting to track, respectively constructing different template sets for different characteristics of a target which needs to be tracked; modeling for the template set of each characteristic by one sparse represented task to construct a multitask combined sparse representation model on the basis of a local sparse graph; carrying out iteration solution on the multitask model by adopting an approximate accelerated neighbor gradient algorithm; selecting weights of different tasks by using a variance ratio; and constructing an appearance model on the basis of multitask combined sparse representation and estimating an optimal state of the target, which is used as a tracking result, by adopting a particle filter algorithm.