33 results about "Anisotropic filtering" patented technology

The invention relates to the field of image processing, in particular to a circuit board microsection image defect detection method, which comprises the following steps: acquiring a circuit board microsection image, calculating local direction energy of each pixel point in the image in a preset scratch direction, and generating a direction energy diagram; calculating a direction continuity index of each pixel point by analyzing neighborhood energy distribution of each pixel point in the preset scratch direction; according to the direction continuity index, constructing a spatially adaptive scratch similarity for each pixel point; and on the basis of the scratch similarity, executing spatial adaptive anisotropic filtering processing on the circuit board microsection image so as to realize accurate detection of circuit board defects. According to the method, by intelligently recognizing the structural difference between the scratches and the defects, the background scratches are accurately removed, meanwhile, the contour integrity of the real defects is kept, and the accuracy and reliability of microsection defect detection are remarkably improved.

The invention discloses an optical coherence tomography detection method for aero-engine thermal growth oxides, and relates to the field of optical detection. The method comprises the following steps: acquiring interference signals of a reference light field and a sample scattering light field, and performing adaptive polarization state compensation on the interference signals based on a dynamic polarization matching factor to obtain compensated interference signals; anisotropic filtering is carried out on the compensation interference signal, oxide layer columnar crystal structure features are separated, and a filtering signal is obtained; performing time-frequency joint analysis on the filtering signal, optimizing a time-frequency transformation kernel parameter by using an adaptive window width adjustment factor, and extracting oxide layer thickness and frequency band energy distribution; and fusing the columnar crystal structure characteristics and the frequency band energy distribution, constructing a cross-scale thermodynamic inversion model, and outputting a prediction result of the residual life of the thermal growth oxide layer of the engine. By introducing polarization compensation, frequency domain analysis optimization and a multi-field coupling inversion model, the imaging definition, the thickness extraction precision and the residual life prediction capability of the thermal growth oxide layer are remarkably improved.

A multi-beam charged particle system and a method of operating a multi-beam charged particle system can provide improved image contrast. The multi-beam charged particle system comprises a filter element or an active array element in a detection system, which can provide improved, anisotropic image contrast. The disclosure can be applied for applications of multi-beam charged particle system, where higher desired beam uniformity and throughput may be relevant.

When sampling a 3D texture using anisotropic filtering, an anisotropy direction along which to take samples in the texture is determined by determining reduced precision representations of the texture coordinate derivative vectors and using the reduced precision texture coordinate derivative vectors to determine a pair of vectors representing the directions of x and y axes for a 2D coordinate system on the plane in the 3D texture defined by the texture coordinate derivative vectors. The x and y axis vectors are used together with the texture coordinate derivative vectors to determine both a X-axis component and a Y-axis component for projected representations of the texture coordinate derivative vectors in the 2D coordinate system on the plane in the 3D texture defined by the texture coordinate derivative vectors. The projected representations of the texture coordinate derivative vectors are then used to determine the anisotropy direction.

The invention relates to the technical field of medical image processing, in particular to a dynamic ultrasonic noise reduction model training method and a dynamic ultrasonic noise reduction method and device. The method comprises the following steps: acquiring an ultrasonic video covering multiple clinical scenes such as antral pneumatosis, hydrops, solid contents, long antral antrum and cardiac anatomical occlusion; marking the precise boundary of the organ frame by frame by a clinical expert, segmenting and complementing a missing section, optimizing the contour of a fuzzy region, and inhibiting or erasing artifacts; a deep learning model is trained based on the labeled data, so that the model can call an exclusive noise reduction algorithm according to organ categories (the antral sinuses adopt morphological reconstruction filtering guided by ellipse / long axis prior, and the heart adopts time-space joint deblurring driven by a cardiac cycle and combining motion compensation and anisotropic filtering); and weighted noise reduction is carried out on the suspected lesion or the user indication area. In the reasoning stage, a high-fidelity and low-noise video stream can be output in real time by inputting any dynamic ultrasonic video, and organ boundary metadata is attached.

A method of performing anisotropic texture filtering involves performing isotropic filtering at each sampling point of a set of sampling points in an ellipse to produce isotropic filter results. Weights of an anisotropic filter are selected that minimize a cost function that penalises high frequencies in the filter response of the anisotropic filter under a constraint that the variance of the anisotropic filter is related to an anisotropic ratio squared, the anisotropic ratio being the ratio of a major radius of the ellipse to be sampled and a minor radius of the ellipse to be sampled. The plurality of isotropic filter results are combined using the selected weights of the anisotropic filter to generate at least a portion of a filter result.

The invention relates to a multi-target detection method, which comprises the following steps of: based on multidirectional anisotropic filtering, effectively decoupling track main body texture and environmental background noise, and remarkably improving the feature expression capability and the signal-to-noise ratio in a low-contrast and complex texture environment; the fracture area of the track caused under the severe working conditions of partial shielding, light reflection and the like is positioned and repaired, and the connectivity and integrity of the output track on the topological level are guaranteed; by constructing a feasible region mask surrounding an orbit, spatial semantic ambiguity caused by a monocular vision perspective effect is eliminated, it is ensured that foreign matter alarm is strictly limited in a physical dangerous region, and the false alarm rate is greatly reduced. Through simultaneous extraction of the railway track topological structure and troubleshooting of the foreign matter risk in the track area, the detection precision under the complex working condition is significantly improved.

This invention discloses a method and system for detecting defects on highly reflective metal surfaces based on industrial vision, relating to the field of industrial vision. The method includes: scanning along the longitudinal direction of a rolling linear guide at a preset speed using a line-scan imaging unit to acquire an original polarized image sequence covering the polarization angle corresponding to the sliding surface of the guide; analyzing the original polarized image sequence, extracting the texture direction field, performing directional anisotropic filtering on the original polarized image sequence, and outputting an enhanced image sequence; using a preset sliding window to progressively capture local image blocks pixel by pixel, and using a scratch probability assessment model to output the probability value of the center pixel belonging to a minor scratch, generating a probability distribution map; performing connected component analysis, and determining continuous pixel regions with probability values ​​exceeding a dynamic activation threshold as minor scratch regions, thereby obtaining the minor scratch defect detection result. This method solves the problems of high false negative rates in highly reflective environments, the inability of fixed thresholds to adapt to reflective environments, resulting in extremely low detection accuracy and strong imaging interference.

The invention is entitled "Graphics Texture Mapping". When performing anisotropic filtering when sampling a texture to provide output sample texture values used in rendering output in a graphics processing system, the number of positions at which the texture is to be sampled along an anisotropic direction in the texture along which samples are to be taken in the texture is determined by determining the square root of the coefficients F of an ellipse of the form Ax 2 + Bxy + Cy 2 = F, the ellipse corresponding to a projection onto a surface to which the texture is to be applied of a sample point at which the texture is sampled, and using the determined square root of the coefficients F of the ellipse to determine the number of positions at which samples should be taken in the texture along the anisotropic direction.

This invention provides an image background noise suppression method based on anisotropic filtering with local contrast, belonging to the field of image processing technology. The method includes: Step S1, calculating the average gray value of pixels in a central neighborhood within a first specific range centered on the current pixel, obtaining the average gray value of the central neighborhood; Step S2, calculating the average gray value of background pixels in eight surrounding background neighborhoods within a second specific range centered on the current pixel, obtaining the average gray value of each of the eight background neighborhoods; Step S3, calculating the differences between the average gray value of the eight background neighborhoods and the average gray value of the central neighborhood in four directions, based on the average gray value of the central neighborhood and the average gray value of the eight background neighborhoods; Step S4, determining the gray value of the current pixel based on the differences in the four directions. This invention provides stronger local signal-to-noise ratio gain, clearer target boundaries, lower average background brightness, and faster calculation speed, enabling it to respond to high frame rate processing requirements.

The application discloses a two-stage SAR image edge detection method based on saliency, and belongs to the technical field of image processing. Anisotropic filtering is used to smooth the input image, so as to filter out noise and retain edge information as much as possible. A Gauss difference operator is used to highlight edge regions and suppress flat regions; a Gabor filter bank and maximum processing are used to generate an edge saliency map, on the saliency map, edge regions have higher saliency values, and other regions are suppressed more thoroughly. Non-maximum suppression and double-threshold processing are used to perform edge rough extraction from the SME, so as to obtain primary edges containing false alarms. A neighborhood discrimination strategy is used to eliminate false alarms from the primary edges, so as to obtain a final edge detection result. Therefore, the application has good scalability, and the detection result has excellent accuracy.

The invention relates to the technical field of image data processing, in particular to a numerical control machining platform remote monitoring method for a high-voltage switch shell, which comprises the following steps: acquiring a machining scene image of a numerical control machining platform containing the high-voltage switch shell; segmenting the processing scene image into a plurality of sub-regions, and determining the disorder degree of each sub-region; determining the attention degree of each sub-region; determining a heat conduction coefficient of each sub-region; based on the heat conduction coefficient, utilizing an anisotropic filtering algorithm to obtain a filtered image of the processing scene image; and based on the filtered image, defect detection is carried out on the high-voltage switch shell by using a target detection technology, so that remote monitoring of the numerical control machining platform of the high-voltage switch shell is realized. The defect that parameters of a traditional anisotropic filtering algorithm are fixed is overcome, dynamic balance is achieved between noise suppression and feature keeping, and the accuracy and reliability of defect detection of the high-voltage switch shell are remarkably improved.

This invention discloses a medical image segmentation method for urinary stones based on region connectivity, belonging to the field of medical image processing technology, to solve the problems of high-density bone interference and topological fragmentation in stone segmentation. First, a piecewise weighted mapping function is constructed to adaptively adjust the window width and perform anisotropic filtering on the image, generating a bone suppression feature map. Then, a maximum stable extremum region algorithm combined with anatomical priors is used to extract high-confidence seed points. A fast traversal algorithm is executed using a constructed gradient resistance field to detect geodesic temporal field abrupt changes to define boundaries, generating a simply connected mask. Finally, three-dimensional dense morphology and texture consistency features are fused for weighted discrimination to remove vascular calcification artifacts. This invention ensures the topological connectivity and high accuracy of the segmentation results.

The invention discloses a self-adaptive anisotropic filtering method, device and equipment and a storage medium, and relates to the technical field of signal processing and filtering, and the method comprises the steps: obtaining a warp gradient and a weft gradient of unfiltered gridding land water reserve abnormal data; analyzing the spatial distribution characteristics of the stripe noise based on the warp gradient and the weft gradient, and dynamically determining the filtering parameters of the anisotropic Gaussian filtering kernel; constructing an anisotropic Gaussian filtering kernel according to the filtering parameters; performing edge filling processing on the unfiltered data to generate a filled expansion matrix; and on the basis of an anisotropic Gaussian filtering kernel and filtering parameters, weighted convolution calculation is carried out on the filled expansion matrix to obtain filtered land water reserve abnormal data, so that strip noise is effectively suppressed, and the intensity of an original signal is reserved.

The invention discloses an image semantic segmentation method and system for wheat seedlings in a tillering stage. The method comprises the following steps: S1, collecting a plurality of first images; s2, carrying out preprocessing based on color correction, ultra-green segmentation and an anisotropic filtering algorithm, and carrying out labeling and fine adjustment on each mask pattern; s3, a segmentation network model is constructed by taking Segform as a trunk, and training is carried out; s4, calculating each evaluation index; and S5, correcting each spectral image and acquiring each phenotypic parameter through feature matching and mask resampling. The system is used for supporting the implementation of the method. According to the method, preprocessing is improved based on color correction, ultra-green segmentation and improved anisotropic diffusion filtering, the image quality is improved, and the accuracy of subsequent segmentation is ensured; an improved segmentation network model is used and a plurality of modules are additionally arranged, so that the feature extraction capability is improved; and background interference is eliminated and a plurality of phenotypic parameters are calculated through various transformations and feature matching, so that the evaluation result is more comprehensive.

The invention discloses an intelligent surveying and mapping analysis platform for multi-source geographic information data fusion, and relates to the technical field of surveying and mapping analysis. Calculating the distribution probability based on a sliding window and quantifying the uncertainty; standardizing the data and then constructing a joint histogram, and calculating a cross-data-source dependency relationship; constructing a convex optimization model containing local entropy and global mutual information, and iteratively solving a spatial adaptive weight; combining a weight matrix and mean filtering to eliminate fusion mutation; non-local mean filtering is adopted to reinforce the edge, and a radial basis function is utilized to fill a data missing area. According to the method, the convex optimization model containing the local entropy constraint and the global mutual information constraint is constructed, so that the adaptive compensation of the spatial heterogeneity of the multi-source data is realized, and a more accurate fusion result is realized in a complex geographic scene.

Sampling a 3D texture using anisotropic filtering, for positions in a render output, comprising: for the corresponding position in the coordinate system of the 3D texture, determining texture coordinate derivative vectors, the derivative vectors defining a plane in the 3D texture; reducing the precision of the derivative vectors; using the reduced precision derivative vectors to determine x and y axis vectors for a 2D coordinate system on the plane; projecting the derivative vectors into the 2D coordinate system; using the projected vectors to determine an anisotropy direction in the 2D coordinate system along which to sample the texture; transforming the anisotropy direction back into the 3D texture coordinate system; sampling the 3D texture along the anisotropy direction; using the samples to determine an output sampled texture value for the positions in the render output. Also disclosed: determining whether both derivative vectors have a same component that is zero; if so using the other components directly for projection to the plane. Derivative vector precision may be reduced from floating point format to signed integer with 10 bit precision. Levels of detail may be determined using mipmaps.

This invention relates to the field of image data processing technology and discloses an adaptive enhancement method for image data in complex environments at the end of a high-speed railway platform. The method includes: acquiring an environmental image and extracting luminance components; performing spatial sampling to construct a luminance distribution matrix; calculating the luminance gradient vector of the sampling nodes to establish spatial weighting coefficients; using the spatial weighting coefficients to perform spatial filtering on the initial mapping gain to generate weighted components and achieve gain correction; calculating the luminance change rate to establish a temporal smoothing step size; performing temporal constraint adjustment on the mapping function control parameters to establish a cooperative mapping function; and using the cooperative mapping function to process the luminance components, identify low signal-to-noise ratio regions, and perform anisotropic filtering. This invention, through spatial weight association and temporal smoothing constraints, eliminates the slice-like effect caused by strong directional light sources, suppresses video flicker, and locks the topological features of the track edge while filtering out noise.

The invention provides a forest tree number and position automatic detection method based on an HDBSCAN algorithm, and the method comprises the steps: carrying out the multi-level threshold separation and denoising operation of original point cloud data, and obtaining forest tree point cloud data; performing initial forest recognition on the forest point cloud data through a density clustering algorithm after dynamic parameter optimization to obtain an initial recognition result; performing three-dimensional shape correction and adjacent target fusion on the initial recognition result to obtain a prediction recognition result; and performing multi-dimensional feature verification on the prediction identification result to obtain a final identification result. According to the method, a forest tree detection task is accurately and efficiently completed through multi-level threshold separation, an anisotropic filtering technology and multi-dimensional verification, and the detection precision and the anti-interference capability are improved.

An apparatus, system, and computer program product for automatically switching between quality modes of an application based on a signal from the application. The apparatus includes an application renderer configured to render an application in a first quality mode. The application renderer is further configured to switch from rendering the application in the first quality mode to rendering the application in the second quality mode when a signal is detected from the application. The signal may be a rendering mode switching message provided by the application, detected from an image frame of the application, or detected from information (e.g., audio) received from the application. The signal may indicate a rendering setting of a second quality mode. Rendering settings may include active shaders, texture resolution, geometric detail levels, MIP map levels, anisotropic filtering, up-sampling levels, anti-aliasing, ray tracing, and global illumination.

The present invention discloses a method and system for image semantic segmentation of wheat seedlings in the tillering stage, the method comprising the steps of: S1, collecting multiple first images; S2, pre-processing based on color correction, super-green segmentation and anisotropic filtering algorithms, annotating and fine-tuning each mask image; S3, constructing a segmentation network model with Segformer as the backbone and training it; S4, calculating each evaluation index; S5, correcting each spectral image and obtaining each phenotypic parameter through feature matching and mask resampling. The system is used to support the implementation of the above method. The present invention improves the image quality and ensures the accuracy of subsequent segmentation by improving the pre-processing based on color correction, super-green segmentation and improved anisotropic diffusion filtering; also uses an improved segmentation network model and adds multiple modules to improve the feature extraction capability; and eliminates background interference and calculates multiple phenotypic parameters through multiple transformations and feature matching, making the evaluation results more comprehensive.

The invention discloses a two-segment SAR image edge detection method based on saliency, and belongs to the technical field of image processing. Anisotropic filtering is adopted to smooth an input image, so that edge information is reserved as far as possible while noise is filtered out. Highlighting an edge region by using a Gaussian difference operator, and inhibiting a flat region; an edge saliency map is generated by using a Gabor filter bank and maximization processing, an edge region on the saliency map has a relatively high saliency value, and other regions are relatively thoroughly inhibited. Performing edge coarse extraction from the SME by using non-maximum suppression and double-threshold processing to obtain a primary edge containing a false alarm; and eliminating false alarms from the primary edge by adopting a neighborhood identification strategy to obtain a final edge detection result. Therefore, the method has good expansibility, and the detection result has excellent accuracy.

The application relates to the technical field of integrated circuit design, in particular to a resource-friendly anisotropic filtering system. The system performs lightweight format conversion on an initial anisotropic ratio and an initial long axis length to obtain a reference anisotropic ratio and a reference long axis length, so that the storage bandwidth, operation units and other resources of a GPU chip are as low as possible under the premise of meeting the anisotropic filtering level calculation accuracy, thereby reducing the consumption of hardware resources of the GPU chip and improving the efficiency of anisotropic filtering level calculation.

A technique for rendering is provided. The technique includes generating a first gradient for a shade space texture tile, wherein the first gradient reflects a relationship between shade space texel spacing and screen space pixel spacing; generating a second gradient for a shade space texel of the shade space texture tile, wherein the second gradient reflects a relationship between material texel spacing and shade space texel spacing; combining the first gradient and the second gradient to obtain a third gradient; and performing anisotropic filtering on the material texture using the third gradient to obtain a value for the shade space texel.

The application relates to the field of image processing, in particular to a circuit board metallographic section image defect detection method, which comprises the following steps: acquiring a circuit board metallographic section image, calculating local direction energy of each pixel point in a preset scratch direction in the image, and generating a direction energy graph; by analyzing the neighborhood energy distribution of each pixel point in the preset scratch direction, the direction continuity index of each pixel point is calculated; according to the direction continuity index, a spatial self-adaptive scratch similarity is constructed for each pixel point; based on the scratch similarity, spatial self-adaptive anisotropic filtering processing is performed on the circuit board metallographic section image, so as to realize accurate detection of circuit board defects. The application realizes accurate removal of background scratches by intelligently identifying the structural differences between scratches and defects, meanwhile, the outline integrity of real defects is preserved, and the accuracy and reliability of metallographic section defect detection are significantly improved.

The present invention discloses an optical coherence tomography detection method for thermally grown oxides in aircraft engines, relating to the field of optical detection. The method comprises: collecting interference signals between a reference light field and a scattered light field of a sample; performing adaptive polarization state compensation on the interference signals based on a dynamic polarization matching factor to obtain a compensated interference signal; performing anisotropic filtering on the compensated interference signals to separate the columnar crystal structure characteristics of the oxide layer and obtain a filtered signal; performing time-frequency joint analysis on the filtered signals, optimizing the time-frequency transform kernel parameters using an adaptive window width adjustment factor, and extracting the oxide layer thickness and frequency band energy distribution; fusing the columnar crystal structure characteristics with the frequency band energy distribution to construct a cross-scale thermodynamic inversion model and output a prediction result for the remaining life of the engine's thermally grown oxide layer. By introducing polarization compensation, frequency domain analysis optimization, and a multi-field coupled inversion model, the imaging clarity, thickness extraction accuracy, and remaining life prediction capability of the thermally grown oxide layer are significantly improved.

When sampling a 3D texture using anisotropic filtering, an anisotropy direction along which to take samples in the texture is determined by determining reduced precision representations of the texture coordinate derivative vectors and using the reduced precision texture coordinate derivative vectors to determine a pair of vectors representing the directions of x and y axes for a 2D coordinate system on the plane in the 3D texture defined by the texture coordinate derivative vectors. The x and y axis vectors are used together with the texture coordinate derivative vectors to determine both a X-axis component and a Y-axis component for projected representations of the texture coordinate derivative vectors in the 2D coordinate system on the plane in the 3D texture defined by the texture coordinate derivative vectors. The projected representations of the texture coordinate derivative vectors are then used to determine the anisotropy direction.

The present application relates to the technical field of image data processing, and particularly relates to a remote monitoring method for numerical control machining platform of high-voltage switch shell, comprising: acquiring a machining scene image of the numerical control machining platform containing the high-voltage switch shell; segmenting the machining scene image into a plurality of sub-regions, and determining the disorder degree of each sub-region; determining the attention degree of each sub-region; determining the thermal conductivity coefficient of each sub-region; based on the thermal conductivity coefficient, acquiring a filtered image of the machining scene image by using an anisotropic filtering algorithm; based on the filtered image, using target detection technology to detect defects of the high-voltage switch shell, so as to realize remote monitoring of the numerical control machining platform of the high-voltage switch shell. The present application overcomes the fixed parameter defect of the traditional anisotropic filtering algorithm, realizes dynamic balance between noise suppression and feature preservation, and significantly improves the accuracy and reliability of the high-voltage switch shell defect detection.

The present application relates to the technical field of image segmentation, in particular to an automatic segmentation system for colorectal tumor CT images. The present application firstly constructs a three-dimensional model of the intestinal cavity, and obtains all flat areas and wrinkle areas, as well as the wrinkle intestinal wall area in each CT image; then, according to the local curvature information of each flat area, the flat area is filtered; in each CT image, the wrinkle extension direction is determined according to the distribution of each wrinkle intestinal wall area to perform anisotropic filtering on the wrinkle intestinal wall area; finally, based on all the filtering results, the three-dimensional model of the intestinal cavity is reconstructed, and the colorectal tumor is automatically segmented by means of a preset segmentation network. The present application is aimed at the structural characteristics of the colorectum, and different filtering methods are used for the flat areas and the wrinkle areas respectively, so that the noise is reduced while the wrinkle texture features of the colorectum are retained to the greatest extent, the over-smoothing of the intestinal wrinkle boundary and the tumor boundary is avoided, and the automatic segmentation effect of the colorectal tumor is improved.

Sampling a texture using anisotropic filtering, comprising: for a texture position for which a sampled texture value is to be provided, determining a base number of sampling positions for which to sam