187 results about "Image gradient" patented technology

The invention relates to the technical field of medical image analysis, in particular to a CAR-T cell culture monitoring system based on image recognition, which comprises a cell region division module, a cell morphology analysis module, a cell behavior dynamic analysis module and a cell population collaborative monitoring module. According to the method, through cell image gradient intensity analysis, refined cell edge detection and dynamic adjustment of a segmentation threshold value, high-precision extraction of cell boundaries is ensured, noise interference is avoided, cell shapes, textures and geometric features are refined, the identification degree of morphological features is improved, and tiny changes of cell morphologies are accurately captured; dynamic tracking and trend evaluation are carried out on cell behaviors based on adjacent frame images, tiny dynamic changes such as cell division, aggregation and migration are accurately reflected, excessive smoothness in the dynamic process is avoided, cell population behavior monitoring is combined with interaction and corresponding speed calculation, the population synergistic effect change trend is comprehensively evaluated, and the cell population behavior monitoring effect is improved. And finer cell culture environment optimization and clinical research support are provided.

The invention relates to the technical field of image processing, in particular to a super-resolution reconstruction system and method based on a pulse neural network, and the system comprises an image pulse encoder, a pulse feature enhancer and a differentiable pulse decoder. An image pulse encoder simulates a receptive field of a retina by using DoG response, performs adaptive pulse distribution on an input low-resolution image in combination with an image gradient, and converts the low-resolution image into a space-time pulse sequence reflecting a high-frequency region and a low-frequency region in the low-resolution image; a pulse feature intensifier performs coarse-grained and fine-grained structure reconstruction of a low-resolution image on the space-time pulse sequence by using a pulse time sequence dependent plasticity mechanism to obtain a granularity feature pulse; and the differentiable pulse decoder converts the granularity characteristic pulse to obtain a corresponding high-resolution image. According to the method, efficient and low-consumption image super-resolution reconstruction is realized through a bionic retina coding mechanism and a pulse time sequence optimization strategy.

The invention discloses a multi-modal remote sensing image matching method, particularly relates to the technical field of remote sensing image processing, and is used for solving the problem of multi-modal image matching. The method mainly comprises the following steps of: 1, improving a phase consistency model, and constructing a phase-moment weighted joint direction feature in combination with a maximum moment and a minimum moment to replace the feature expression of the traditional image gradient; 2, implementing a point product fusion strategy on the phase-amplitude characteristics extracted by the phase consistency model and the maximum moment, and constructing phase-moment weighted joint amplitude characteristics; and 3, on the basis of the steps 1 and 2, identifying the direction of the feature points and screening local peak values to determine the main direction. Three values adjacent to a peak value are selected, and the peak value position is interpolated through parabola fitting so as to improve the matching precision; and step 4, constructing a logarithm polar coordinate descriptor based on regularization non-uniform partition to generate a feature description vector. Through the mode, high-precision and high-efficiency matching of the multi-mode remote sensing image can be realized.

The invention relates to the technical field of image precision optimization, in particular to a surveying and mapping image mathematical precision evaluation method based on dense point cloud matching, which comprises the following steps: acquiring and preprocessing image data, synthesizing an image gradient modulus length and a maximum principal curvature, and introducing a surface normal vector partial derivative to obtain a surface normal vector partial derivative; calculating a curvature-driven depth discretization step length, and generating an initial point cloud data set; constructing a point cloud complex network structure diagram and calculating the topology durability of a point cloud topology structure to obtain an optimized point cloud data set; based on the optimized point cloud data set, in combination with the target image information, the adaptive curvature gradient weight and the error diffusion control value, constructing an adaptive matching cost function, and calculating a point cloud matching cost matrix; and calculating an optimal point cloud transformation matrix based on the point cloud matching cost matrix. According to the surveying and mapping image mathematical precision evaluation method based on dense point cloud matching, low-texture region point cloud matching error suppression and structure optimization of high-curvature region point cloud on a complex curved surface are realized.

The invention discloses a digital imaging method of an ear-nose-throat examination endoscope, and relates to the technical field of medical treatment, and the method comprises the following steps: synchronously collecting a white light image, a photoacoustic signal, a stimulated Raman spectrum, pressure sensor data and a physiological signal of a target orifice through a multi-physical field probe; constructing a four-dimensional tensor bound with the anatomical features; based on a vocal cord vibration fundamental frequency harmonic characteristic optimization graph convolutional network, dynamically constructing an adjacent matrix and coupling cross-modal characteristics to generate a submucosal lesion enhanced image; a deformable convolutional network constrained by vocal cord biomechanics is adopted, and the shape of a convolution kernel is dynamically adjusted according to the relation between real-time strain and elastic modulus, so that motion artifacts caused by swallowing actions are inhibited; generating a curvature-driven asymmetric convolution kernel based on ear canal spiral geometry, and executing super-resolution reconstruction in combination with confrontation training of fractal constraint; and fusing the white light image gradient and the pressure gradient field, and outputting a three-dimensional lesion contour consistent with the anatomical structure.

The invention provides a steel rail seam width detection method and system based on line-line distance, and the method comprises the steps: shooting an image of a rail connection part, obtaining a near-end rail and a far-end rail through an image recognition strategy, defining the boundary points of the top edge of the near-end rail as A and B, and defining the boundary points of the bottom edge of the near-end rail as C and D, the B and the C are located on the same track connecting line, the far-end track is defined in the same mode, and the A point and the B point are obtained; according to the C point determination strategy, determining the C point based on multi-factor weighting, and obtaining a near-end key point set; acquiring a far-end key point set according to the image acquisition processing step; two sets of key lines of the near-end track and the far-end track are obtained respectively, and the distance between the two sets of key lines is calculated to serve as the width of the steel rail seam; through multi-level constraints of image gradient features and track theoretical width, the problem that a single image feature is susceptible to noise interference is solved, and the positioning precision of A and B points is improved.

The invention discloses an image quality evaluation method based on dynamic weight feature extraction and three-dimensional fusion under a meta-learning framework, and the method comprises the steps: designing a multi-receptive-field and dynamic parameter module in an image quality evaluation model, extracting feature information under different views, and enriching and weighing multi-receptive-field features; screening out a salient region by using image gradient information outside the image quality evaluation model, designing a distortion classification model based on the salient region, and providing shared features for the image quality evaluation model; using random cutting and three-dimensional convolution to deeply fuse the multi-receptive-field features and the shared features to obtain deep fusion features; shartcut Vision Transform is proposed to perform semantic analysis on the fusion features, the association between the overall features and quality fluctuation is summarized, and the quality score of the image is obtained through linear mapping. The multi-distortion image quality evaluation method can accurately evaluate the quality of the multi-distortion image, can effectively process a large amount of data, and has the advantages of being objective, rapid and high in usability.

The invention discloses a reversible information hiding method and system for enhancing image smoothness, and relates to the technical field of information security, and the system comprises a preprocessing module, an information embedding module and an optimization recovery module. According to the method, the Sobel operator is utilized to calculate the image gradient map, the region is divided based on the threshold T, the smooth region and the texture region can be accurately distinguished, an accurate basis is provided for differential processing of different regions, region characteristics are fitted, large-size blocks and sub-blocks of the smooth region are divided to facilitate subsequent refinement processing based on the gradient mean value, and the processing efficiency is improved. The small-size blocks of the texture area are beneficial to operation aiming at texture features, the rationality and effectiveness of overall processing are improved, meanwhile, the pixel mean value in the large-size blocks of the smooth area is calculated, the pixel and mean value difference value is recorded, reversible low-pass filtering is carried out, and the reversibility of the image in the whole information hiding and recovering process is ensured.

A method for displaying and processing an image includes capturing the image, calculating one or more of a size, a position, and a field of view of a first region of interest (ROI) of the captured image, calculating one or more of a size, a position, and a field of view of a second ROI of the captured image, calculating a blur factor for blurring the first ROI, manipulating an image gradient of the first ROI pixel by pixel using the blur factor to reduce at least one of a brightness or a contrast of one or more pixels of the ROI, manipulating the second ROI by changing a color of one or more pixel values from the second ROI, and displaying the processed image on a display. An imaging system which includes an image signal processor (ISP) is also described.

The invention discloses an image processing method and system for extrusion shaping of a wood-plastic plate, and the method comprises the steps: carrying out the multi-scale analysis of an image of the wood-plastic plate, and recognizing the local features in the image; on the basis of multi-scale image feature extraction, a deformation mode in an image sequence is recognized through a time sequence image data association and dynamic correction method, and image processing parameters are adjusted in real time; according to the dynamically corrected image, processing the image by using morphological operation, and extracting a surface defect area of the wood-plastic plate; after morphological operation, identifying a defect area based on an image gradient optimization method; and on the basis of the sharpened image, analyzing the local flatness in the image by using a region growing algorithm. According to the method, the feature points are extracted from the image sequence and matched, the deformation mode between the images is recognized, and the local surface flatness is accurately analyzed in combination with the region growing algorithm, so that accurate detection of the surface defects and irregular regions of the wood-plastic plate is realized.

The disclosure deals with system and method for image gradient and derivative computation image processing. Noise, image sharpness, orientation, empirical parameters, and computational complexity are examples of image gradient and derivative computation challenges. Many traditional kernel-based operators excel at tackling one of these problems, but trade off their ability to handle others. Two new gradient detection kernels based on two-dimensional high order Taylor Series expansion tackle many such problems. The first kernel uses a wide range of the pixels in view to suppress noise, thereby improving the gradient intensities of edges. The second kernel builds on the first to leverage its noise suppression benefits while tackling an additional problem of degraded and low contrast edge boundaries. It can detect smooth lines in the presence of discontinuities and poor quality. The filter architecture allows for precise gradient calculation, edge detection, and orientation determination to less than one degree of the true value even when faced with signal to noise ratios that exceed 0.75.

The invention provides a fusion method for an infrared image and a visible light image, and belongs to the technical field of computer vision and artificial intelligence, and the method comprises the steps: S1, constructing an image fusion network model for generating a fusion image; s2, acquiring the infrared image and the corresponding visible light image to form a training data set, and training the fusion network model by using a loss function by using the training data set, the loss function is related to the structural similarity loss of the visible light image and the fusion image, the structural similarity loss of a semantic thermodynamic diagram mask area based on the infrared image and the corresponding visible light image, the image intensity loss and the image gradient loss; and S3, inputting an infrared image to be fused and a corresponding visible light image into the fusion network model trained in the step S2 to obtain a final fusion image. According to the method, manual operation is not needed, the computing power requirement is lowered, real-time image processing can be achieved, information is not prone to being lost, and efficiency is higher.

The invention relates to the technical field of defect detection, in particular to a corrugated paper defect detection method and system based on machine vision. The method comprises the following steps: acquiring a corrugated paper image and converting the corrugated paper image into a frequency domain; identifying a periodic texture peak point based on amplitude spectrum angle energy distribution, generating a texture suppression filter, and combining with a multi-scale Gaussian high-pass filter to obtain a composite filter bank; the filter bank is used for filtering the frequency domain signals and performing energy weighted fusion; performing inverse transformation on the fused signal to obtain an enhanced image; and segmenting the enhanced image to obtain candidate defects, and verifying by combining the gradient information of the original image to determine a final defect. According to the scheme, the texture background can be adaptively inhibited, the multi-scale defect is enhanced, the false alarm rate is effectively reduced in combination with gradient check, and the detection accuracy and robustness are improved.

The invention discloses a visual calibration method, a terminal and a computer storage medium. The method comprises the following steps: generating theoretical grid coordinates; controlling a motion platform carrying a visual imaging unit to move to each target moving position in sequence, wherein the positions are obtained by superposing random two-dimensional offsets in a preset range; at each target moving position, a laser galvanometer and a visual imaging unit are synchronously triggered through the same hardware trigger source, and single-point mark engraving and image acquisition are executed; calculating a weighted centroid of the mark based on an image gradient magnitude, and extracting a sub-pixel center coordinate; combining the visual internal reference, the fixed coordinate transformation relation and the geometric constraint of the bearing plane, and resolving actually measured physical coordinates; carrying out space alignment, deviation calculation and invalid point interpolation on theoretical and actually measured coordinates, and generating a distortion data file which can be identified by a controller; the method further comprises a retry and fault-tolerant processing mechanism when coordinate calculation fails. According to the scheme, the calibration precision, the robustness and the automation level are effectively improved.

The invention relates to the technical field of coating materials and detection, in particular to a surface defect detection method and system for a silicon carbide coating graphite disc. Performing multi-dimensional image data acquisition on the surface of the graphite disc; noise removal is carried out through an adaptive multi-scale convolutional neural network, and detail and edge information is retained in combination with image gradient constraints; high-frequency details are recovered by using an image enhancement technology, geometric correction is performed through a deep learning spatial transformation network, and data consistency is ensured; an authentication authority sequence is introduced, so that the authenticity and the generality of image data are ensured; an improved YOLOv9 model is adopted, an attention mechanism is introduced to enhance the extraction capability of target features, a Soft-NMS algorithm is combined to optimize a positioning frame, the overlapping detection problem is solved, and a traditional convolutional layer is replaced with an SCM module; a detection result is fed back in real time, a detailed detection report is generated, and decision support is provided. The surface defect detection precision and efficiency of the silicon carbide coating graphite disc are improved.

The invention relates to the field of image processing, in particular to a quality monitoring method and system for wire harness processing, and the method comprises the steps: obtaining a wire harness gray-scale map; for each pixel point of the grey-scale map, comparing a gradient histogram of a neighborhood of the pixel point with a preset template image gradient histogram, and finding out suspected matching points; calculating a gradient direction deviation angle between the suspected matching point and the template image, rotating a neighborhood of the suspected matching point based on the deviation angle, calculating a gray value of the suspected matching point after rotation, repeatedly rotating each pixel point in the neighborhood of the suspected matching point and calculating the gray value, and generating a corrected neighborhood; and matching the correction neighborhood of the pixel point with a preset template image, calculating a correlation coefficient between the correction neighborhood of the pixel point and the preset template image, and carrying out wiring harness abnormal region detection based on the correlation coefficient and a preset abnormal threshold value. According to the invention, by generating the correction neighborhood which is more fit with the template, the false detection and omission ratio can be effectively reduced, and the reliability of the detection result is improved.

The invention relates to the technical field of image recognition control, in particular to a linear motion module accurate control method and system based on AI vision, and the method comprises the following steps: obtaining an interference fringe image and extracting a gray path, analyzing central point displacement to generate a coordinate sequence, repairing an interference feature position, and aligning the image and a control time sequence to generate a synchronization time table. The mapping structure response forms a set of control output configurations. According to the method, a peak path is constructed through interference fringe gray scale information in an image frame, central point pixel displacement is extracted, a coordinate migration sequence is constructed, fine-grained capture and tracking of an execution state are realized, an external interference area is identified according to image gradient change, vector repair is carried out, and image data stability and feature reliability are guaranteed. And aligning the repaired image response with a control instruction time sequence, constructing an impulse response lag statistical mechanism, realizing accurate mapping adjustment of the control frequency, and improving the control precision and response consistency in high-frequency displacement execution.

The invention discloses a microscopic visual threshold segmentation algorithm based on OTSU improvement for a micro-nano object image, and the algorithm comprises the steps: S1, combining image gradient information, calculating a new gray scale weight factor based on an index and a logarithmic function, and carrying out the bilateral filtering of an image; s2, calculating a traditional gray threshold value of the filtered image by using a traditional between-class variance formula, and calculating an improved gray threshold value by using an improved between-class variance formula based on a natural constant function; and S3, carrying out adaptive iterative operation on the pixels of the image by using the traditional gray threshold and the improved gray threshold so as to complete three-region threshold segmentation. According to the OTSU-based improved micro-nano microscopic visual image threshold segmentation algorithm disclosed by the invention, the problem of inaccurate calculation result caused by pixel quantity proportion imbalance during threshold segmentation is solved, and the limitation of poor local analysis capability of a global threshold segmentation algorithm is improved.

The invention discloses a desulfurization optimization control method and system based on a multi-modal model, and relates to the technical field of industrial intelligent control, and the method comprises the steps: constructing a gas-liquid interface boundary point set, and generating a control feature vector; calculating fractal dimensions of control feature vectors by using a box counting method, constructing a disturbance point set, multiplying disturbance intensity by an expansion trend to obtain an unstable diffusion grade, matching key propagation points with nearest disturbance points by using nearest neighbor matching, and generating a control priority sequence; and defining a control unstable diffusion level, defining an optimization objective function, and solving by using constraint optimization to obtain a final control variable. According to the method, a composite strength factor is constructed through an image gradient field and a signal response ratio, a gas-liquid interface boundary point set and a control feature vector are constructed, the robustness and information density of features are improved, a main resonance frequency is recognized by combining a fractal interference field and short-time Fourier transform, key propagation points are screened by using wavelet transform and information entropy analysis, and a high-precision gas-liquid interface is obtained. The desulfurization efficiency and the stability of the reaction tower are improved.

The invention relates to a double-branch gradient information guided dark light image enhancement method, which comprises the following steps of: firstly, extracting low-illumination image gradient information by using a Sobel operator; a shallow network feature used for optimizing global brightness balance, a middle network feature used for correcting local area brightness of the image and a deep network feature used for carrying out contrast adjustment of scene perception in high-level semantic features are sequentially obtained through a depth feature enhancement branch; according to the image features after gradient information extraction, gradient features of a plurality of stages are obtained in sequence in combination with a gating gradient optimization module; and obtaining an enhanced image through an attention fusion module. According to the method, the problems of high-frequency detail loss and insufficient real scene generalization ability in a low-light environment in an existing method can be solved. Meanwhile, the enhanced image processed by the image enhancement deep network guided by the double-branch gradient information is not easy to have an overexposure phenomenon, and the feature description force of the network on a dark light image is effectively improved.

The invention provides a multi-mode aerospace remote sensing cooperative surveying and mapping regional network refinement method. The method comprises the following steps: constructing an aviation optics and SAR image rigorous positioning model under a geocentric rectangular coordinate system; extracting a rational polynomial model parameter RPC under a satellite optics and SAR image geocentric rectangular coordinate system; constructing a positioning model and a distance equation of a satellite laser height measurement remote sensing target; extracting the position of the light spot energy center on the light spot image, and enhancing the definition of the light spot image; homonymy point matching of a multi-mode remote sensing target is achieved through image gradient feature matching, homonymy point extraction between optical images is achieved through sift feature or least square matching, and matching between laser height measurement data and remote sensing images is achieved through light spot energy center position extraction and template matching; extracting homonymy points between the SAR and the Lidar data by using an SAR simulation image matching technology; constructing a multi-modal remote sensing area network joint adjustment model; and integrally solving the refined parameters, and solving the coordinates of the encryption points.

The invention relates to the field of subway fault analysis, in particular to a subway fault data analysis method based on a cloud platform, and the method comprises the steps: obtaining train operation difference parameters and noise characteristic difference parameters of all key monitoring points, so as to determine a monitoring difference state; determining a difference analysis strategy of each key monitoring point position according to the monitoring difference state; when abnormal defect analysis is carried out, a defect analysis mode is determined according to the noise period stability coefficient and the image gradient difference coefficient of the key monitoring point position; under the condition that abnormal defect analysis is completed, a monitoring compensation mode is determined according to the defect interference radiation coefficient of the key monitoring point position; when parameter optimization analysis is carried out, the setting mode of the point location anomaly coefficient is determined according to the noise anomaly frequency of the key monitoring point location and the operation interference proportion, whether the point location anomaly coefficient is adjusted or not is determined according to the coincident anomaly parameters, and the accuracy of the operation parameter optimization result is improved.

The invention relates to the technical field of Internet of Things and remote monitoring, in particular to an unattended platform monitoring and alarming system and method based on an Internet platform. The method comprises the following steps of: acquiring space parameters of a platform, and performing sound field signal acquisition to obtain a sound pressure level numerical value group; sound source space positioning is carried out according to the sound pressure level numerical value group, and sound source space coordinate data and a sound pressure level space-time distribution diagram are obtained; deploying a camera array according to the platform space parameters, and performing platform scene image analysis to obtain an image frame sequence, enhanced image data, a multi-scale image pyramid and image gradient data; performing optical flow equation solving according to the multi-scale image pyramid and the image gradient data to obtain sparse optical flow data; and generating a motion vector field according to the sparse optical flow data. According to the invention, the accuracy and reliability of platform monitoring are greatly improved through multi-modal fusion perception, refined anomaly detection and intelligent analysis and decision.

The invention relates to the technical field of medical image processing, in particular to an acupuncture guide system based on image registration. According to the method, the height parameter field of the human body surface three-dimensional model is established and the contour line density distribution is extracted, so that hierarchical division and density aggregation identification of morphological characteristics of a target area can be realized, and local geometric characteristic identification is carried out by carrying out tangential analysis on direction changes of continuous contour line segments and combining Gaussian curvature. According to the method, geometric abnormal point locations of the surface structure can be accurately calibrated, and an area with direction stability and geometric continuity can be discriminated in multi-point data in combination with a space consistency judgment mechanism in the normal direction, so that the space identification process is more robust, and the accuracy of the space identification is improved. Multi-feature re-check judgment of a fitting end point can be realized through density peak value and normal consistency dual screening, the spatial accuracy of a positioning point is guaranteed, and the consistency of structural continuity and visual guidance information can be considered through a path construction mode generated by combining an image registration result and image gradient information.

The invention relates to the technical field of oral decayed tooth detection, and discloses an oral decayed tooth detection system and detection method. Through quadric surface fitting and smoothing processing, image noise is reduced and local details are reserved. By calculating an image gradient and constructing a gradient field, the gray level change direction and amplitude are effectively extracted. The divergence calculation and the weighted aggregation degree improve the accuracy of abnormal region identification, and especially in early-stage decayed tooth detection, the abnormal region can be highlighted and the normal region can be ignored. The self-adaptive threshold setting is combined with the local mean value and the standard deviation, the threshold is dynamically adjusted, and false detection and missing detection are reduced. And 8, gathering the scattered candidate points into a complete area according to an adjacency principle and connectivity analysis, and ensuring that a decayed tooth area with an irregular shape is not missed. The method ensures that the final detection area has clinical significance through the screening of features such as physical area and boundary pixel number, and enhances the operability of the detection result through red contour labeling.

The invention discloses an image binarization method based on a dynamic window. The method comprises the steps of 1, target image noise suppression and gray level conversion; 2, carrying out compression processing on the dynamic range of the bright part of the image through a logarithmic function; 3, calculating a global threshold value of the grayscale image by utilizing a maximum between-class variance method; 4, dividing the grayscale image into image blocks according to the image gradient, and calculating a local threshold value for each block by using a Sauvola algorithm; 5, according to the global threshold value and the local threshold value of each image block, the threshold value of each image block is obtained through calculation according to the weight proportion; and 6, segmenting each image block by using a corresponding threshold value to obtain a final segmented image. According to the invention, the size of a local window is adaptively changed when any pixel point is binarized; and the image binarization capability under the condition of non-uniform illumination in a complex illumination environment is remarkably improved.

The present invention relates to the field of image processing technology, and more specifically to a cross-view heterogeneous image matching method for airborne sensors. The method comprises obtaining a registration image and a reference image to be matched; extracting edge features of the registration image and the reference image respectively; coarsely aligning the registration image and the reference image; matching based on the coarsely aligned positions to obtain feature pairs, and organizing them into a multi-scale feature pair set; and estimating the transformation matrix between the images using a RANSAC algorithm based on the multi-scale feature pair set, ultimately completing precise image matching. The present invention effectively achieves a balance between matching accuracy and computational speed by combining a mechanism of fast coarse registration based on image gradient edge features with precise multi-scale feature registration based on image phase information. Furthermore, the method exhibits good robustness for airborne heterogeneous images.

The present application belongs to the technical field of mobile robot navigation, and particularly relates to a SLAM navigation method and system for a mobile intelligent cabinet, which comprises the following steps: acquiring a laser point cloud sequence and a grayscale image of the mobile intelligent cabinet and performing data cleaning; obtaining a geometric feature index based on the spatial jump distribution of the laser point cloud sequence; obtaining a visual texture index based on the local dispersion of the image gradient; calculating laser dynamic weight and visual dynamic weight by using the geometric feature index and the visual texture index, weighting and fusing the pose change quantity calculated by the single-line laser radar odometry and the visual odometry to obtain a fused pose quantity; and updating the global state based on the fused pose quantity and driving autonomous navigation. The present application can adjust the sensor weight in real time according to the environmental characteristics, solves the problem of positioning divergence in the long corridor of a shopping mall or a high-reflectivity environment, and improves the robustness of navigation.

The invention discloses a multi-angle template matching method based on gradient topology sparse coding. The multi-angle template matching method comprises the following steps of S1, extracting a gradient topology skeleton of a template image; s2, constructing and coding a skeleton sparse dictionary; s3, coefficient domain rotation mapping construction is carried out; s4, performing multi-angle matching and similarity calculation; and S5, result fusion and optimization. The invention relates to the technical field of computer vision and image processing, and has the beneficial effects that the features based on the gradient topology skeleton are insensitive to illumination variation, gray shift and noise interference can be effectively overcome, and the matching stability in a complex environment is improved. By constructing the coefficient domain rotation mapping relation, physical rotation reconstruction does not need to be carried out on the image, the calculation overhead of multi-angle matching is remarkably reduced, the search efficiency is improved, sparse coding is used for carrying out compact representation on the image structure, and the adaptive capacity to local deformation and partial shielding is enhanced.

The invention discloses an endoscope video frame specular reflection highlight detection method and system, and belongs to the field of endoscope optical detection. The method comprises the following steps: firstly, converting an endoscope video frame into a grey-scale map; a Sobel operator is utilized to extract edge gradient to generate a binary image gradient mask, and meanwhile, a binary image highlight mask is generated by obtaining an area with the brightness obviously higher than the average level of a grey-scale map; dividing the binary image gradient mask and the binary image highlight mask into image blocks, adaptively adjusting the sizes of the blocks according to the highlight ratio of the image blocks so as to fuse the binary image gradient mask and the binary image highlight mask, and performing morphological processing and region screening to obtain a final binary image highlight mask of the grey-scale image; and carrying out time domain compensation on the final binary image highlight mask to obtain a specular reflection highlight detection result of the endoscope video frame. The accuracy, integrity and robustness of specular reflection highlight detection of the endoscope video frame are effectively improved.