193 results about "Pixel mapping" patented technology

The invention discloses a video style migration attitude detection method based on a graph neural network, which belongs to the technical field of video processing, and specifically comprises the following steps: preprocessing a video data stream; constructing a dynamic space-time association network, mapping pixels of the video frame into space-time diagram nodes, and converting the original video frame into a target artistic style by aggregating the space-time features of the nodes layer by layer; in the style migration process, a dynamic mask is generated for the area where the human body joint points are located, and the geometric consistency of the joint structures in stylized output is restrained; extracting joint point coordinates of each frame from the stylized video, constructing a space-time trajectory sequence, and generating a feature vector containing space-time dependence; performing multi-dimensional comparison on the feature vector and a preset action mode library, calculating a track offset and a mode matching degree, judging that the action is abnormal according to a threshold value, and outputting a detection result; according to the invention, the cooperative performance of style migration and anomaly detection is improved.

The invention relates to the technical field of data processing, and discloses a visual data comprehensive processing method for automobile industrial data based on digital twinning. The method comprises the following steps: performing recursive quartering on a three-dimensional space of an automobile general assembly workshop through a workshop process pixel quadtree index algorithm, performing process type marking on multi-source heterogeneous data according to a process identification mapping rule, and allocating a workshop real-time data set to a quadtree index structure according to space coordinates to perform process state aggregation, and according to the viewport pixel mapping relation, performing weighted aggregation calculation on the process state aggregation matrix to obtain a pixel comprehensive display value, and based on the pixel comprehensive display value, performing layered rendering processing on the workshop multi-process parallel scene. According to the invention, the visual processing efficiency and the user interaction response speed of the digital twin system in the automobile industry are improved.

The invention discloses a deburring method combining a 2D image and a 3D point cloud, and relates to the technical field of image processing, and the method comprises the steps: carrying out the point cloud registration under the verification of the surface orientation consistency of a standard point cloud and an actual point cloud; constructing a pixel mapping matrix between the workpiece image and the RGB image; semantic-level edge coarse extraction is carried out on the workpiece image, and extraction of a sub-pixel-level actual edge contour is carried out on the basis of coarse extraction in combination with a traditional edge detection algorithm; aligning the actual edge contour to the actual point cloud through the pixel mapping matrix; and the target polishing surfaces are aligned to the actual point cloud, to-be-polished edge contours are screened based on the distances between the contours of all the target polishing surfaces in the actual point cloud and all the actual edge contours, and a polishing track is obtained in the workpiece image. According to the method, the sub-pixel edge positioning capability of the 2D image and the spatial topology information of the 3D point cloud are fused, and the inherent defect of a single mode is overcome.

The invention relates to the technical field of video recognition and analysis, and discloses a video recognition and analysis method based on physical characteristics. Video frame pixels are mapped to a two-dimensional coordinate system with the upper left corner as an original point, and mirror image expansion and median filtering are carried out on a gray level image; constructing a binary image based on a gray threshold value, and analyzing and extracting a target region by using a four-neighborhood connected domain; using neighborhood search and polar angle sorting to close the tracking contour, and generating equidistant re-sampling points based on Euclidean distance and an interpolation method; the curvature of the re-sampling points is estimated through a three-point difference algorithm, and zero denominator is avoided through numerical protection; performing discrete Fourier transform on the curvature sequence to extract a frequency spectrum, and normalizing an amplitude to form a standardized feature vector; and finally, inter-frame similarity is calculated based on the feature vector, and the most similar frame is automatically retrieved. By processing unified data standards in stages, edge noise is suppressed, sampling uniformity is ensured, and feature stability and cross-frame comparability are improved.

The invention discloses a reinforcing steel bar surface binding point identification and positioning method based on point cloud and pixel mapping. The method comprises the following steps: obtaining original three-dimensional point cloud data to carry out outlier noise point elimination; performing plane fitting on the denoised point clouds, identifying template surface point clouds and removing the template surface point clouds; mapping the spatial physical coordinate attribute of the steel bar surface point cloud into a three-channel coding image; separating the three-channel coded image into three single-channel images: a length image, a width image and a depth image; processing the depth image to extract an initial reinforcing steel bar straight line set; screening out a parallel steel bar straight line set from the initial steel bar straight line set according to a set reference direction and a parallel angle tolerance threshold value; and calculating straight line intersection points in the parallel steel bar straight line set to obtain candidate binding point pixel coordinates, determining representative binding point pixel positions through clustering, querying coordinate information of the three single-channel images, and outputting three-dimensional space coordinates of the steel bar binding points. According to the invention, robust and accurate steel bar binding point identification and positioning are realized.

The invention relates to the technical field of machine vision detection and intelligent image processing, and discloses an intelligent track defect identification method based on machine vision. The track center is automatically positioned through a pixel coordinate system with the upper left corner as the original point in combination with row gray scale accumulation and maximum projection, the region of interest is dynamically intercepted, and background noise and calculation amount are effectively reduced. For the area, multi-angle equal-interval projection is adopted, pixels are mapped into a one-dimensional sequence, and the response to various strips and slender defects is improved. Each projection sequence is subjected to secondary trend fitting, a comprehensive residual image is constructed through the maximum residual, and defect signals are enhanced. According to the scheme, a threshold value is determined in a self-adaptive mode through statistics, four-neighborhood connected domain extraction and a dynamic area threshold value are combined, and a noise small domain is removed. Boundary, main direction, length and severity features are extracted from the remaining connected domains and mapped back to an original image, and efficient, accurate and automatic track defect intelligent recognition is achieved.

The invention discloses a skin sensitivity identification method and system based on a conditional generative adversarial network, and the method comprises the steps: constructing a pixel-to-pixel mapping relation data set according to a preset image mapping relation; inputting the collected real skin original image into a generative network structure Unet, and generating a false image with the same size as the skin original image; forming an image pair by the generated false image, the real skin original image and the real skin effect image, and inputting the image pair into a discriminator network PatchGAN to obtain a true and false image probability; calculating the loss Loss according to the output of the generator and the discriminator, and updating the network parameters of the generator and the discriminator by adopting a gradient descent method to obtain an image migration model; inputting a to-be-detected skin image into the image migration model to obtain a skin sensitive effect picture, and calculating a sensitive region area of the skin sensitive effect picture; and designing a skin sensitivity score mapping relation according to the area of the sensitive region, and evaluating the skin health state. The real-time performance is high, and the quality of the generated image is high.

The invention relates to the technical field of endoscopes, and provides a spectral image processing method for a three-dimensional endoscope. The method comprises the following steps: synchronously acquiring single-band spectral images returned by multiple sensors of the endoscope; after the single-band spectral image is corrected, a two-dimensional spectral image is generated by combining multi-level fusion of the spectral weight; constructing a three-dimensional point cloud coordinate set through parallax calculation; mapping the two-dimensional spectrum fusion image to a three-dimensional point cloud based on a back projection sub-pixel mapping algorithm to obtain a three-dimensional spectrum point cloud; and performing Poisson fusion driven inter-block splicing fusion on the three-dimensional spectrum point cloud, and outputting an interactive three-dimensional navigation model. The technical problem that image details are lost and three-dimensional reconstruction is inaccurate due to the fact that spectral image fusion precision is insufficient in an existing three-dimensional endoscope image processing method is solved, registration and reconstruction of high-precision three-dimensional spectral point clouds are achieved through multi-modal image collaborative correction and optimization fusion, and the image fusion precision is improved. And the accuracy and the real-time performance of navigation in the endoscope are improved.

The invention relates to the technical field of target tracking, in particular to a panoramic target tracking method and device based on fusion positioning, equipment and a storage medium. Comprising the steps of obtaining a panoramic video stream and a positioning coordinate sequence; performing pixel mapping on each positioning coordinate in the positioning coordinate sequence to obtain a current mapping coordinate and a historical mapping coordinate; performing delay compensation on the current mapping coordinates based on the historical mapping coordinates to obtain predicted mapping coordinates; performing target detection on the panoramic video stream to obtain a visual coordinate set of the current frame, and performing spatial matching on the predicted mapping coordinate and each visual coordinate in the visual coordinate set to form a matching candidate set; if the matching candidate set is null, outputting the predicted mapping coordinate as a target position coordinate; and if the matching candidate set comprises at least one candidate visual coordinate, outputting the candidate visual coordinate closest to the predicted mapping coordinate as a target position coordinate. According to the invention, the tracking performance under a complex environment condition can be improved.

The invention relates to a forest region satellite fire point monitoring method based on deep learning, and the method comprises the steps: collecting a multispectral image of a target forest region through a satellite remote sensing system, and carrying out the preprocessing of the multispectral image; feature information extraction is carried out on the preprocessed multispectral image, the extracted feature information is classified by using a convolutional neural network, and a classification feature vector is obtained; fire source areas are determined according to the classification feature vectors, and a first fire source area set is obtained; acquiring environment data, and acquiring a second fire source region set through a support vector machine model according to the environment data and the preprocessed multispectral image; according to the first fire source area set and the second fire source area set, real fire point pixels are obtained; and real fire point pixels are mapped to geographic coordinates of the target forest region, missing data are filled by adopting a spatial interpolation technology, and final fire point position distribution is obtained. According to the method, the accuracy and reliability of forest fire detection under the complex terrain condition are remarkably improved.

The invention relates to a light effect animation data generation method and apparatus, a device and a medium. The method comprises the steps of obtaining a target image for constructing light effect animation data; according to the color values of the pixels in the target image, constructing a palette containing a fixed number of reference colors, and determining attribution probability distribution corresponding to each reference color mapped to the palette by each pixel; corresponding to the parity of each pixel in the plane space position of the target image, selecting a reference color according to the attribution probability distribution so as to update the color value of the pixel; and packaging the target image with the updated color value into the light effect animation data. According to the method, the light effect animation data is generated by utilizing the attribution probability and odd-even spatial modulation, and the smooth pixelated light effect can be realized when the atmosphere lamp is played.

The invention relates to the technical field of image detection, in particular to a power battery flatness detection method based on image recognition. The method comprises the following steps: acquiring a power battery image and calibration board information, and determining a pixel mapping relation; generating a battery surface point cloud picture by using the pixel mapping relation; fitting a reference plane by using the battery surface point cloud atlas, and calculating point cloud deviation; marking a battery bump area according to the point cloud deviation; calculating the height deviation of the bulge area of the battery, and determining the bulge degree; marking surface sunken points based on the battery surface point cloud atlas, and calculating the sunken depth of the surface sunken points; when the recess depth is greater than the recess threshold value, marking as a recess abnormal region; the flatness of the power battery is evaluated according to the bulging degree and the sunken abnormal area; and converting the flatness of the power battery into a three-dimensional rendering graph, and outputting a flat area grading result. According to the invention, accurate quantification of the battery flatness is realized based on an image detection technology, and the flatness detection accuracy is improved.

A single-frame speckle deflection measurement method based on subarea wavefront correction belongs to the technical field of optical measurement, and comprises the following steps: based on a monocular deflection measurement system, projecting a standard speckle image to a to-be-measured mirror surface from a screen, and collecting a single-frame deformed speckle image reflected by the to-be-measured mirror surface; processing the reflected image by using a Zernike polynomial projection distortion correction method based on a mark point, and correcting the projection distortion of the reflected image caused by a non-orthogonal light path; obtaining sub-pixel mapping of the reflection image and the standard image based on a sub-region wavefront fitting correction and iterative registration algorithm; and based on system calibration and camera calibration, establishing a spatial incident light model and a spatial reflected light model, obtaining surface gradient distribution of the to-be-measured mirror surface, and reconstructing a three-dimensional shape of the to-be-measured mirror surface. According to the method, a reference plane is not needed, high-reflection surface shape measurement of only collecting a single-frame image can be achieved, the dynamic range and precision of speckle deflection measurement are improved through subarea wavefront correction, and the surface reconstruction precision can reach hundreds of nanometers.

The invention relates to the technical field of image generation, in particular to a medical beauty 3D model construction method and system based on an image large model, and the method comprises the following steps: obtaining rasterized image data based on a 3D full-head point cloud model through space coordinate projection and pixel mapping, judging connectivity and a missing region, screening region features, and generating texture guidance parameters; and screening textures according to partition features, determining a multi-region fitting index, and outputting a continuous boundary fusion region. According to the invention, through carrying out multi-angle spatial reconstruction and fine pixel mapping on the three-dimensional point cloud data, texture coverage is not limited to a single view angle, spatial feature partitioning and target attribute identification, so that accurate supplementation of the texture of the missing region is realized; the locally generated texture content and the original input region are highly fused in the aspects of color structure and detail representation, and region mutation is effectively eliminated through spatial adaptation and boundary adaptive processing, smooth connection of multi-part texture splicing presentation, color transition and gradient fusion.

The invention provides a virtual camera projection method and system, electronic equipment and a medium, and belongs to the technical field of target detection, and the method comprises the steps: constructing a training data set according to sampling points in a three-dimensional space; modeling projection and back projection between the real camera view and the virtual camera view to obtain a two-stage multi-layer perceptron model; training a two-stage multi-layer perceptron model according to the training data set to obtain a trained model; and obtaining a pixel mapping table between the virtual camera view and the real camera view based on the trained model. According to the method, the reversible pixel mapping relation between the real camera and the virtual camera is constructed by adopting the two-stage multi-layer perceptron model, and the pictures under any group of camera parameters are uniformly converted to be under the virtual camera parameters, so that high-quality label data are efficiently generated through the automatic labeling model; the data utilization rate and the generalization ability of the automatic labeling model are remarkably improved, so that the model can directly process multi-source heterogeneous data, and the model adaptation cost is reduced.

A multi-camera imaging system includes a plurality of imaging units arranged side-by-side to capture images of a scene. A calibration module is configured to determine intrinsic and extrinsic parameters of camera modules of the imaging units and to establish a per-pixel mapping from image coordinates to a three-dimensional space. An image processing module selects, for individual of the plurality of imaging units, a field-of-view region within a captured image and blends the fields-of-view from adjacent imaging units together to form a composite image.

The invention relates to the technical field of intelligent monitoring, in particular to an embedded multi-lens panoramic video intelligent monitoring system and a construction method thereof, a computing energy SE5 micro server and an NVIDIA Jetson board card construction system are integrated, a deep learning neural network is adapted to the computing energy SE5 micro server, and the NVIDIA Jetson board card construction system is adapted to the NVIDIA Jetson board card construction system. The functions of network model weight conversion, sample generation, deployment training, test analysis and the like are completed; embedded spherical panoramic splicing software for lookup table calculation and GPU parallel acceleration is adapted on a Jetson board card platform to complete lookup table import, pixel mapping, suture line fusion and CUDA deployment and optimization functions, and meanwhile, a YOLO-based vehicle illegal parking detection, retrograde driving detection and human-vehicle statistics module is designed according to the parking lot monitoring requirements of the highway service area, so that the parking lot monitoring requirements of the highway service area are met. According to the method, the real-time splicing and intelligent analysis processing functions of the high-resolution panoramic video are realized on a heterogeneous embedded platform through collaborative optimization, and the application bottleneck of OMS panoramic video splicing and intelligent analysis processing on the embedded platform at present is solved.

A multi-camera imaging system includes a plurality of imaging units arranged side-by-side to capture images of a scene. A calibration module is configured to determine intrinsic and extrinsic parameters of camera modules of the imaging units and to establish a per-pixel mapping from image coordinates to a three-dimensional space. An image processing module selects, for individual of the plurality of imaging units, a field-of-view region within a captured image and blends the fields-of-view from adjacent imaging units together to form a composite image.

The invention provides a spectral measurement method of a two-dimensional spectrometer, which comprises the following steps of: forming a two-dimensional spectrogram by adopting a cross dispersion mode, and calculating the spectrogram to obtain spectral information of a measured light source; the method comprises the following steps: S1, calculating an ideal spectrogram model: obtaining a corresponding relation between a dispersion wavelength and a spatial position of incident light under an ideal condition by utilizing a dispersion equation of a cross dispersion element and structural parameters of an optical system; s2, obtaining a corrected spectrogram model: measuring a light source with known characteristic spectral lines to obtain a two-dimensional spectrogram, and performing parameter inversion on the structural parameters of the optical system involved in the step S1 to obtain the corrected spectrogram model; s3, establishing wavelength-pixel mapping: performing coordinate value fitting on wavelength data in the corrected spectrogram model to complete mapping from the corrected spectrogram model to actual pixel coordinates; and S4, measuring the spectral information of the measured light source: obtaining the spectral information of the measured light source by using the wavelength-pixel mapping relationship established in S3, thereby effectively improving the accuracy, stability and efficiency of the spectral measurement of the two-dimensional spectrometer.

The present invention relates to a calibration measurement method and system based on pixel mapping. The method comprises: providing a camera for measurement, and performing pixel mapping-based calibration on the camera, wherein, during calibration, a pixel position-horizontal resolution mapping relationship of the camera is established, and a pixel horizontal offset fitting line of the image captured by the camera is determined; when measuring an object based on the calibrated camera, at the current pixel coordinates of the object image, the spatial height of the object and the corresponding width change in the horizontal direction of the image are determined based on the camera's pixel position-horizontal resolution mapping relationship and the pixel horizontal offset fitting line; based on the determined spatial height of the object and the corresponding width change in the horizontal direction of the image, single-frame contour information is spliced along the object scanning direction and step size to output a three-dimensional point cloud of the object. The present invention can achieve rapid calibration of the camera, and when measuring with the calibrated camera, a high-precision three-dimensional point cloud can be output.

The invention relates to the technical field of optical waveguide preparation, and discloses an optical waveguide mother set splicing-free preparation method based on a DMD dynamic mask, and the method comprises the steps: obtaining the layout data of an optical waveguide mother set as initial design information, and generating a mask pattern sequence which comprises a plurality of partition mask patterns; establishing a pixel mapping relationship between the on-off state of the micro-mirror unit and the partition mask pattern; driving the transparent substrate to execute partition positioning movement on the platform, and controlling the current exposure area of the photosensitive material layer to be aligned with the projection area of the partition mask pattern; on the basis of the pixel mapping relation and the partition positioning movement result, activating the mask pattern, and carrying out non-interval continuous exposure to form a full-page latent image pattern; the whole latent image pattern is subjected to developing treatment, the embossment structure of the optical waveguide master mask is treated and transferred into the transparent substrate, a waveguide core layer pattern is formed, the effects of non-interval continuous exposure and no segmented splicing are achieved, the preparation period of the surface embossment optical waveguide master mask is shortened, and the processing efficiency and the batch preparation yield are improved.

A calibration system for multiple image sensors includes a manufacturing apparatus, a first imaging sensor, a second imaging sensor positioned at a different location than the first imaging sensor and configured to capture at least one different wavelength than the first imaging sensor, a calibration artifact, and an imaging calibration module. The imaging calibration module is configured to obtain first and second images of the calibration artifact via the first and second imaging sensors, determine first and second pixel mappings between the first and second images and a common coordinate system, according to a location of the calibration artifact in the first and second images, convert an image of a manufacturing object captured by the first imaging sensor to the common coordinate system, and convert an image of the manufacturing object captured by the second imaging sensor to the common coordinate system.

The invention relates to the technical field of computer vision and image processing, in particular to a space target three-dimensional reconstruction method based on epipolar line mapping constraint dense search. The processing can reduce the dependence on transformation matrix projection parameters in the weak texture region matching process. A kernel line coordinate system is established, orientation processing is carried out on the relative direction of a matched feature image pair, a new inter-pixel mapping relation is obtained, the inter-pixel mapping relation is used as dense search constraint in a triangulation network, and dense search has a better search effect in a weak texture area. And integrating the kernel line mapping search result and the existing epipolar constraint relationship, and finally performing multi-view reconstruction to obtain a target physical model. According to the method, dense point searching between matched image pairs can be effectively carried out, pixel point level searching precision can be achieved, and three-dimensional reconstruction of a space target is achieved.

The application relates to the technical field of display, and particularly discloses an extended reality head-mounted display system based on Micro LED display, which comprises an eye movement tracking device, an image rendering processing unit, a bionic pixel mapping driving unit, a Micro LED display panel and an event-driven control unit; the line-of-sight center position is acquired through eye movement tracking, non-uniform rendering and bionic pixel mapping are realized, full-resolution pixels are driven in the line-of-sight center area, the data amount is reduced by using a pixel merging algorithm in the peripheral area, and the refresh is triggered only in the picture change area in combination with an event-driven mechanism. The application significantly reduces the computing power and bandwidth demand, relieves visual fatigue and dizziness, simultaneously realizes dynamic power consumption optimization, and improves the immersive experience and equipment endurance.

According to one embodiment, a processing apparatus for accelerating convolution operations comprises a memory array in which a plurality of elementary processing unit are arranged in an array structure, and an operation unit configured to temporarily store feature-map data, divide the feature-map data into predetermined units along a channel-axis direction according to a pixel-first mapping method, and deliver the divided data to the memory array in a manner of shifting the data according to a position of a kernel.

The invention discloses a video stitching method and device based on an embedded platform, and the method comprises the steps: collecting a video stream with an overlapping region through two cameras, carrying out the processing of the first frame data on the embedded platform, obtaining a homography matrix, and calculating the overlapping region; determining a suture line by using an optimal suture line algorithm, and performing color correction and fusion on the first frame by using an optimized histogram matching method and a fade-in and fade-out method; detecting a moving object by using a frame difference method, setting a current frame as an updated frame, an updated suture line and a color correction function when the moving object passes through the vicinity of the suture line, avoiding ghosting and blurring, and periodically updating a homography matrix in order to eliminate errors; and generating pixel mapping from the data of the first frame and the updated frame, storing a color correction function, directly performing coordinate mapping and color correction on other frames, accelerating a fusion process by using hardware, and finally displaying a spliced video on a display device in real time. The splicing quality is improved while the real-time performance is ensured, and the spliced image is quickly obtained and displayed on the display device in real time.

This application relates to a method, apparatus, electronic device, and storage medium for calculating road surface height. The method includes: acquiring lane lines; mapping lane line pixels to a first coordinate system to generate first coordinates; acquiring the next time-of-flight (RTK) point; calculating the distance between the next time-of-flight RTK point and the point containing the first coordinates to generate a deflection distance value; acquiring camera height data; processing the deflection distance value and the camera height data; and outputting the road surface height value. The solution provided in this application can replace the traditional method of calculating using the current time-of-flight RTK point and camera height data, thereby improving the accuracy of road surface height and thus improving the accuracy of the three-dimensional spatial coordinates of lane lines.

The present application relates to the technical field of image rendering, in particular to a rendering method and system for a model image interface, comprising the following steps: obtaining a view angle parameter and generating a screen pixel mapping result set of a surface patch, calculating a screen occupation ratio of the surface patch, generating a screen change feature of the surface patch in combination with a frame difference, generating a rendering scheduling sequence set based on the occupation ratio and the change feature, calling data in sequence to perform rasterization processing, and generating a rendering result of the model image interface. In the present application, the visual presentation weight under the current visual field is calculated by extracting vertex coordinate mapping, the difference in occupation ratio between the front and back frames is analyzed by using the finite difference method to capture the dynamic change trend, the dynamic rendering scheduling sequence is constructed based on the double dimensions of the occupation ratio and the change feature, the calculation resources are accurately tilted to the high screen occupation ratio and the high dynamic change area, the rendering quality of the core visual area is ensured, the invalid calculation overhead is greatly reduced, and the interface rendering real-time performance and frame rate stability in complex scenes are effectively improved.

The application provides a video blind area compensation transmission method and server based on dynamic field of view fusion, through acquiring multi-source video stream data, integrating all field of view regions under the same timestamp to a unified global reference space to obtain a global field of view splicing layout, constructing a dynamic field of view topology evolution graph containing a blind area hole outline and a pixel mapping transformation relationship between fields of view according to layout changes of continuous timestamps; performing pixel-by-pixel spatial migration on adjacent field of view video frames according to the pixel mapping transformation relationship of the blind area hole periphery to generate a blind area compensation video segment; performing panorama fusion on the compensation segment and the original video segment to obtain a panoramic video stream; analyzing the distribution pattern of the original collection source and the content remapping source in the panoramic video stream to generate a spatial mask, setting a differentiated encoding transmission strategy for different regions and encoding and distributing. The application realizes high-quality compensation of blind area content and regional differentiated efficient transmission of panoramic video streams.