Patents
Literature
Patsnap Eureka AI that helps you search prior art, draft patents, and assess FTO risks, powered by patent and scientific literature data.

409 results about "Pixel color" patented technology

High-fidelity three-dimensional reconstruction method for mirror reflection plane

A high-fidelity three-dimensional reconstruction method for a specular reflection plane comprises the steps of decomposing pixel colors into diffuse reflection and specular reflection components based on a 3D Gaussian sphere, introducing a dynamic reflection ratio parameter and a spherical harmonic function coefficient to respectively represent two reflection characteristics, and simulating light multi-reflection behaviors through weight fusion of cumulative projection and a reflection ratio map. Secondly, in combination with monocular inverse depth calibration, depth smoothing constraint of color gradient weighting and edge mutual exclusion loss, geometric consistency is enhanced, and artifacts are suppressed; a progressive multi-resolution training strategy is further adopted, low resolution is gradually optimized to complete resolution, reflection parameters are activated in stages, 3D Gaussian sphere overgrowth and floating artifacts are inhibited, and efficiency and precision are balanced. According to the method, the reconstruction fidelity of the specular reflection scene is remarkably improved while the real-time rendering advantage of the 3DGS is reserved, and the method is suitable for the high-precision modeling fields of virtual reality, augmented reality and the like.
Owner:TSINGHUA SHENZHEN INTERNATIONAL GRADUATE SCHOOL

Radiance field gradient scaling for unbiased near-camera training

Methods and systems disclosed herein relate generally to radiance field gradient scaling for unbiased near-camera training. In a method, a processing device accesses an input image of a three-dimensional environment comprising a plurality of pixels, each pixel comprising a pixel color. The processing device determines a camera location based on the input image and a ray from the camera location in a direction of a pixel. The processing device integrates sampled information from a volumetric representation along the ray from the camera location to obtain an integrated color. The processing device trains a machine learning model configured to predict a density and a color, comprising minimizing a loss function using a scaling factor that is determined based on a distance between the camera location and a point along the ray. The processing device outputs the trained ML model for use in rendering an output image.
Owner:ADOBE INC

Nerve radiation field rendering method and device based on nerve base and tensor decomposition

The invention discloses a neural radiation field rendering method and device based on a neural base and tensor decomposition, and belongs to the technical field of computer vision, and the method comprises the steps: collecting a multi-view image and camera parameters thereof to construct a data set; the method comprises the following steps: sampling a radiation field subjected to tensor decomposition according to camera parameters under a selected training view angle to obtain a density feature tensor and a texture feature tensor, respectively converting the density feature tensor and the texture feature tensor into a density value and a neural base, calculating an affine matrix and bias through a coefficient network based on the neural base, and performing affine transformation in combination with an observation direction to obtain a coefficient; inputting the neural bases and the coefficients into a neural hybrid network to obtain a color of a single sampling point, calculating a pixel color by using a volume rendering technology, and training the whole network; and setting a new visual angle, obtaining corresponding camera parameters, and performing image rendering by using the trained network parameters. According to the method, a high-quality three-dimensional scene can be efficiently reconstructed from multi-view image data, a complex surface reflection effect is effectively reconstructed, and vivid image rendering under a new view is realized.
Owner:ZHEJIANG JUNTONG INTELLIGENT TECH CO LTD

Sub-pixel curve rendering in content generation systems and applications

Approaches presented herein provide for the generation of image content including fine objects, such as those that may be less than one pixel in width. Fine objects such as hairs can be represented by curve equations, and curves defined by these equations can be given an inflated width in order to perform conservative hit testing to efficiently identify pixels that a hair might intersect. Once candidate pixels are identified through ray tracing or sampling, for example, false positives or pixels with little hair representation can be removed from consideration. For remaining pixels, a linear representation of the inflated curve can be used to determine intersection points and vertices of a pixel quad corresponding to a hair, which can be used to generate a convex geometric shape representative of the object. The percentage of the pixel area occupied by this geometric shape can be used to determine an alpha or blending value for blending the pixel value of this hair or object with the background pixel color.
Owner:NVIDIA CORP

Generating a rendered image of a three-dimensional object

PCT designated stage expiredWO2024241064A1AnimationImage data processing detailsData packFeature vector
Data representing a 3D object to be rendered is obtained at a user device, the data comprising: a mesh structure defining 3D co-ordinates for each of a plurality of object vertices; and a feature vector encoding visual characteristics for an object surface defined by the plurality of object vertices. An artificial neural network, ANN, is selected from a plurality of ANNs stored on the user device, each of the plurality of ANNs being configured to output pixel colour values for the object surface based on at least some of the visual characteristics encoded in the feature vector, wherein different ANNs have different numbers of layers and / or different numbers of parameters, wherein the selection is based on a resource characteristic of the user device. The mesh structure is processed using the selected ANN to generate a rendered image of the three-dimensional object.
Owner:SONY COMP ENTERTAINMENT EURO LTD

Non-vision field imaging detection method and system of fusion event camera

The invention discloses a non-vision field imaging detection method and system fused with an event camera, and the method comprises the following steps: enabling the event camera and a standard optical camera on a detection prototype to be aligned with a ground region of a corner, forming a video file based on a tiny pixel color difference amplification technology, and carrying out the calculation and analysis of the video file; processing the video file by using an image matching algorithm, and eliminating noise by using Kalman filtering; establishing a linear system, and then obtaining a reconstructed image by using a forward transmission light intensity function; performing Euler video amplification processing on the reconstructed image, wherein the Euler video amplification comprises motion estimation, spatial decomposition, time filtering and signal enhancement; detecting the detection area, and mapping the motion track of the target hidden behind the corner to a two-dimensional plane; the system comprises a simulation scene construction module, a three-dimensional point cloud set construction module and an information integration module. According to the invention, a micro motion algorithm and optical flow calculation are popularized to non-vision field imaging, so that the optical imaging technology can act on a vision field range and also can act on a non-vision field range.
Owner:杭州智元研究院有限公司

Method for traffic sign quality assessment

A method includes receiving image data captured by a traffic feature detection system. The image data is representative of a traffic feature within an environment. The method includes identifying a type of the traffic feature. The method includes isolating a portion of the image data containing the traffic feature. Based on the portion of the image data containing the traffic feature, the method includes determining a pixel color value for the traffic feature. Based on the type of the traffic feature, the method includes determining an expected pixel color value for the traffic feature. Based on a comparison of the expected pixel color value and the determined pixel color value for the traffic feature, the method includes determining a health status for the traffic feature.
Owner:GM GLOBAL TECHNOLOGY OPERATIONS LLC

High frame rate rendering method and system based on display driving chip

The invention relates to a high-frame-rate rendering method and system based on a display driving chip, and the method comprises the following steps: carrying out the bidirectional scattering distribution processing of sorted scene data based on a preset real-time radiance algorithm, and obtaining scattering scene data; performing time domain reprojection on the scattering scene data through a preset frequency domain analysis technology to obtain time domain scene projection data; performing high-frame-rate pixel coloring rendering processing on the time domain scene projection data based on a preset programmable pipeline technology to obtain coloring rendering data; and performing frame synchronization processing on the coloring rendering data through a cache management mechanism to obtain synchronous rendering data, and inputting the synchronous rendering data into a preset display driving chip to drive a display screen to appear a target rendering picture, thereby solving the problem that a traditional method usually depends on a single rendering algorithm or a hardware acceleration scheme, and is not easy to implement. And smooth dynamic picture display cannot be realized on the premise that the image quality is not sacrificed.
Owner:SHENZHEN OSTAR DISPLAY ELECTRONIC CO LTD

Generating a rendered image of a three-dimensional object

PCT designated stage expiredWO2024241063A1AnimationImage data processing detailsMorphingFeature vector
Data representing a 3D object to be rendered is obtained, the data comprising: a mesh structure defining 3D co-ordinates for each of a plurality of object vertices; and a feature vector encoding visual characteristics for at least one object surface defined by the plurality of object vertices. Motion information indicative of motion of the object in a scene is received. Based on the motion information, the mesh structure is deformed by adjusting the co-ordinates for one or more of the plurality of object vertices. The deformed mesh structure is processed using a graphics pipeline comprising a vertex shader and a fragment shader to generate a rendered image of the scene. The fragment shader comprises an artificial neural network trained to output pixel colour values for the at least one object surface on the basis of the visual characteristics encoded in the feature vector.
Owner:SONY COMP ENTERTAINMENT EURO LTD

Method, device and equipment for generating orthoimage based on point cloud and storage medium

The invention relates to the technical field of remote sensing images, in particular to a method, device and equipment for generating an orthoimage based on point cloud and a storage medium. Comprising the following steps: S1, acquiring point cloud data; s2, carrying out grid processing on the point cloud data; s3, sorting single grid units comprising a plurality of points according to elevation values, retaining top point clouds, determining pixel color values of the grid units, generating a three-channel image with a plurality of pixels, and constructing a preliminary ortho-image map; s4, traversing the three-channel image to identify null value pixel points, extracting coordinate values of a plurality of non-null value pixel points in the neighborhood, and calculating a coordinate azimuth angle difference value; and S5, according to the coordinate azimuth angle difference value, judging whether the null value pixel point is a hole pixel point located in the image, when the hole pixel point exists, carrying out image repairing, and outputting a repaired orthoimage. According to the method, the high-fidelity orthoimage map can be quickly generated, and the processing efficiency is remarkably improved while the authenticity and the integrity of the ground feature color are ensured.
Owner:CHINA RAILWAY NO 2 ENG GROUP CO LTD +1

Compensation method and device of display panel and computer readable storage medium

The invention discloses a compensation method and device of a display panel and a computer readable storage medium. The compensation method of the display panel comprises the following steps: acquiring brightness data and chromaticity data of the display panel; determining a brightness compensation coefficient of each sub-pixel color in the to-be-compensated area according to the brightness data of each sub-pixel color in the to-be-compensated area and the brightness data of each sub-pixel color in a brightness reference area; determining a chromaticity compensation coefficient of each sub-pixel color of the to-be-compensated area according to the chromaticity data of the to-be-compensated area and chromaticity data of a chromaticity reference area; determining a total compensation coefficient of each sub-pixel color in the to-be-compensated area according to the brightness compensation coefficient and the chromaticity compensation coefficient; and compensating the to-be-compensated area by using the total compensation coefficient. The display uniformity of the display panel can be improved.
Owner:HEFEI VISIONOX TECH CO LTD

Gait identity information desensitization method and system based on sparse contour hopping

The invention provides a gait identity information desensitization method and system based on sparse contour hopping. The method comprises the steps that S1, a gait video is deframed into an image sequence, and binarization processing is conducted on the image sequence; s2, extracting identity features in the binary image sequence by using a gait recognition model, and calculating an adversarial disturbance matrix of each image according to a loss function; s3, binary image sequence key frames are selected at equal intervals, and adversarial disturbance meeting the pixel jump constraint is added to the human body contour edge in the key frames; s4, embedding the finally added disturbance value into the binary image sequence to obtain a reversible human body contour image sequence; and S5, pixel color backfilling is carried out on the modified binary image sequence, and a desensitized gait video is obtained. According to the method, the loss function of the image sequence before and after desensitization is maximized by means of the gait recognition model, so that the pixel directly generates binary hopping towards the gradient rising direction, non-directional or directional interference on the gait recognition result is realized, and the identity information privacy is protected.
Owner:SHANGHAI JIAOTONG UNIV

Method for identifying irrigation state of plant growth area based on image identification

The invention relates to the technical field of image recognition, in particular to a method for recognizing the irrigation state of a plant growth area based on image recognition, which comprises the following steps: acquiring an image of the plant growth area, segmenting the plant area and extracting contour coordinate points of each leaf; and calculating the mean value of the ratios of the centroid spacing to the perimeter of each contour to obtain a leaf contour space distribution factor, and based on the leaf contour space distribution factor and the segmented plant region pixels. According to the method, by extracting the centroid position and the spatial structure feature of the leaf contour, the spatial distribution difference of each leaf in a plant area is quantified, and further through contour coordinate iteration alignment operation, more stable and unified group leaf shape features are obtained, and interference caused by leaf shape difference is reduced; meanwhile, precise quantification and symbiotic relationship analysis of a pixel color channel are adopted, spatial position information and color differences are combined, and a heterogeneity index of fusion of the leaf shape and color distribution is constructed.
Owner:WUXI UNIV

Digital printing image color detection method

The invention discloses a digital printing image color detection method, and belongs to the field of image data processing and the technical field of image detection. The invention provides a printed image color detection method of an improved superpixel segmentation algorithm based on saliency and color information entropy, and the method comprises the steps: delimiting an important region of a printed image through employing an image saliency algorithm, and adaptively determining the size of a superpixel of the region according to the color information entropy of the printed image in a saliency region. Performing different superpixel division on the salient region and the non-salient region of the printed image; in the printed image color detection, the color mean values of the printed image under the same superpixel label are compared, and the label with chromatic aberration is output, so that a chromatic aberration area is marked in the image, and the color offset of the chromatic aberration area is displayed. The method is used for printing image color detection of superpixel color mean value comparison, and a more accurate color detection result conforming to human vision can be obtained; the method has a fast detection speed, and can output a color difference area position and a color offset.
Owner:ZHEJIANG UNIV OF TECH

Point cloud dyeing method, device and equipment and readable storage medium

The invention discloses a point cloud dyeing method, apparatus and device, and a readable storage medium. The method comprises the steps of obtaining a point cloud frame and a plurality of image frames; aiming at each space point of each camera device, generating an observation transformation matrix used for representing the projection movement condition of the space points on an image plane in the exposure process of the camera devices; generating a relative timestamp for indicating the imaging moment of the spatial point; based on the relative timestamp and the observation transformation matrix, generating a pixel coordinate of the spatial point projected to an image frame of the camera device, and based on the pixel coordinate, performing pixel color picking on the spatial point to obtain a pixel vector; establishing a corresponding relationship between each spatial point of different camera devices and each point cloud in the point cloud frame; and determining a color vector of the corresponding point cloud based on the pixel vector of each spatial point corresponding to the same point cloud. Visibly, according to the method, space-time alignment of the three-dimensional space points and the pixel coordinates on the image frame can be realized by constructing the observation transformation matrix and the relative timestamp, and the point cloud dyeing precision is improved.
Owner:WHITE RHINO ZHIDA (BEIJING) TECH CO LTD

Self-adaptive display method of LCD (liquid crystal display) screen, terminal equipment and storage medium

The invention belongs to the technical field of display, and particularly relates to a self-adaptive display method of an LCD (liquid crystal display) screen, terminal equipment and a storage medium. And according to the frame rate of the video to be played and the pixel color change condition in the color sequence corresponding to each frame, determining that the duration required for completing each color change in the color sequence by the pixel is within the duration limited by the frame rate, if not, carrying out replacement adjustment on the color in the color sequence by using the replacement color in the replacement color set, and if not, carrying out adjustment on the color in the color sequence by using the replacement color in the replacement color set. According to the method and the device, each color change in the pixel completion color sequence is ensured to be within the limited duration, so that the problem of smear and ghosting during video playing is prevented, and the replacement process is not perceived by a user because the replacement color and the replaced color are the same color which cannot be distinguished for the user, so that the overall watching experience of the user is ensured.
Owner:SHENZHEN DESHAN XINTONG ELECTRONICS CO LTD

Three-dimensional reconstruction method for boundary-free scene of extraterrestrial celestial body based on Gaussian radiation field

The invention relates to an extraterrestrial celestial body boundary-free scene three-dimensional reconstruction method based on a Gaussian radiation field, and the method comprises the steps: extracting visual features of an input image, and obtaining internal parameters, external parameters and scene sparse point cloud of a camera through feature matching and cluster adjustment; an expanded dense point cloud is calculated according to a light plane intersection mode, a series of three-dimensional Gaussian distributions are constructed, and pixel colors in a given visual angle direction are rendered in a rasterization mode; aiming at a rendering result, guiding a Gaussian position to be accurately close to a scene surface by utilizing pre-estimated depth prior in an optimization process; a scene surface is accurately captured in combination with an explicit and implicit model joint representation mode; and rebuilding a fine and complete three-dimensional mesh model on the surface of the extraterrestrial celestial body by using a marching tetrahedron algorithm according to an optimization result. According to the method, fine three-dimensional reconstruction of the surface of the extraterrestrial celestial body in a boundary-free scene can be realized, the modeling capability of a Gaussian radiation field in the boundary-free scene is enhanced, and the precision and integrity of three-dimensional reconstruction of the surface of the extraterrestrial celestial body are improved.
Owner:BEIHANG UNIV

Space target three-dimensional reconstruction realization method based on radiation field generation

The invention discloses a space target three-dimensional reconstruction realization method based on radiation field generation. The method comprises the following steps: acquiring image data of a space target for preprocessing; an improved generated radiation field network is built, and an image is sampled to obtain an image patch; obtaining a group of three-dimensional sampling points in a layered sampling mode, performing multi-frequency position coding, splicing with a shape vector, sending into a shape coding layer to obtain feature representation, and inputting into a color head together with a sight direction code and an appearance variable to calculate the color of the sampling points; for each ray, colors of all sampling points are synthesized into a final pixel color based on a volume rendering equation, and the final pixel color is combined into a patch image to be output; inputting the real patch and the patch image into a discriminator, and calculating confrontation loss and discrimination loss; and acting the weight obtained by training on a generator to obtain rotating videos of the space target at different visual angles. The method can realize three-dimensional reconstruction of a space target, and has the advantages of flexible data driving, cross-view geometric consistency, efficient training and the like.
Owner:NANJING UNIV OF AERONAUTICS & ASTRONAUTICS

Color value determination method and apparatus, electronic device, computer storage medium and computer program product

Disclosed in embodiments of the present disclosure are a color value determination method and apparatus, an electronic device, a computer storage medium and a computer program product. The color value determination method comprises: acquiring a shifted color value of each texel amongst a plurality of texels affecting a target pixel, the shifted color value of each texel being obtained by performing first shift processing on a texel color value of the texel on the basis of the texture format of the texel, the numbers of decimal places of decimals of the shifted color values in different texture formats all being the same, and the number of integer digits of said decimals all being the same; on the basis of an impact factor of each texel on the target pixel and the shifted color value of each texel, determining a pixel color value of the target pixel; and according to a restoration mode corresponding to the texture format, restoring the pixel color value so as to obtain a target color value of the target pixel. The present disclosure can reduce the occupied space of hardware modules, thus saving the occupied space for graphics processing units.
Owner:MOORE THREADS TECH CO LTD

Rock debris particle image instance segmentation method under white light condition

PendingCN119027435ASolve the problem of difficulty in accurate instance segmentationimprove accuracyImage enhancementImage analysisPattern recognitionComputer graphics (images)
The invention relates to the technical field of computer vision and image processing, in particular to a rock debris particle image instance segmentation method under a white light condition, and the method comprises the steps: carrying out the superpixel extraction, and carrying out the superpixel segmentation of a rock debris particle image through employing a graph-based efficient image segmentation method; superpixel color similarity merging: performing post-processing on a segmentation result of the image-based efficient image segmentation method by using a color similarity-based superpixel merging method, and merging superpixels belonging to the same rock debris; and superpixel morphological similarity combination: combining superpixels which belong to the same rock debris and have different colors by using a morphological similarity-based superpixel combination method, and completing rock debris particle image instance segmentation under a white light condition. The method has the advantages of being high in accuracy, free of manual data annotation, suitable for various types of rock debris, insensitive to shooting environments such as illumination angles, high in robustness and the like.
Owner:CHINA NAT PETROLEUM CORP +1

Single-pixel color imaging system based on multi-wavelength radio frequency marking

The invention relates to the technical field of optical imaging, and discloses a single-pixel color imaging system based on multi-wavelength radio frequency marking. The system comprises a light source module used for generating three laser beams with three optical primary colors; the radio frequency driving module is used for generating a radio frequency complex frequency signal; the light path adjusting module is used for performing heterodyne interference processing on the three laser beams according to the radio frequency complex frequency signal, respectively forming a light spot array and a reference light source and then combining the beams, and generating a combined beam to be incident on a measured target; each laser beam corresponds to one group of sub light spot arrays; sub light spot arrays with aligned spatial positions exist among the three groups of sub light spot arrays; the detected target detection module is used for collecting light data generated after a detected target reflects the combined light beam; and the image inversion module is used for performing image inversion processing on the optical data to generate a true color image of the measured target. According to the system, the light spot array of multi-wavelength laser incidence can be registered in the spatial position.
Owner:HANGZHOU INST FOR ADVANCED STUDY UCAS

Laser radar and camera joint calibration method, electronic equipment and storage medium

The invention relates to the field of computer vision, and discloses a laser radar and camera joint calibration method, electronic equipment and a storage medium. The method comprises the following steps: acquiring point cloud data of a plurality of laser radars and image data of a camera; performing motion distortion compensation on the point cloud data to generate a global point cloud map; converting the global point cloud map into a plurality of three-dimensional Gaussian voxels with color attributes and covariance attributes; minimizing the first loss function by jointly optimizing the attribute of the three-dimensional Gaussian voxel and the external parameter of the camera so as to complete the calibration of the external parameter of the camera; wherein the first loss function comprises a color consistency constraint term and a rendering error term, and the color consistency constraint term is used for constraining that observed color attributes of a single three-dimensional Gaussian voxel at multiple moments and multiple visual angles are kept consistent; the rendering error term is a pixel color error between a projection image generated by the three-dimensional Gaussian voxel and image data of the camera. Therefore, the precision of external parameter calibration in a dynamic scene is improved.
Owner:江淮前沿技术协同创新中心

Liquid crystal display device, image display method, electronic apparatus, and readable storage medium

The invention discloses a liquid crystal display device, an image display method, electronic equipment and a readable storage medium. In a first working mode, a control circuit of the device receives display data of an mth frame of image, and controls the device to display n sub-field images to obtain the mth frame of image; the display data comprises initial gray scales under n primary colors; in one sub-field image of the mth frame of image, pixels comprise a target pixel, and the actual gray scale of the target pixel is an average value of the middle gray scale and the initial gray scales of T non-target pixels; the actual gray scale of at least part of non-target pixels is the initial gray scale; the target pixel at least meets one of a first condition and a second condition; the first condition is that the distance between a target point reflecting the color of the target pixel in the reference color space and a white point is smaller than a distance threshold value; the second condition is that the reference brightness of the target pixel is greater than a brightness threshold; and / or the contrast ratio of the target pixel is greater than a contrast ratio threshold value.
Owner:BOE TECHNOLOGY GROUP CO LTD

Soil color surveying and identifying system and method

The invention discloses a soil color exploration and identification system and method, belongs to the technical field of soil color identification, can effectively reduce the influence of ambient light on soil color discrimination through a standardized color correction board and software image processing, and performs hierarchical division on a soil color identification image along a vertical plane, thereby improving the accuracy of soil color identification. Pixel point color information of a plurality of independent soil layers is obtained, so that the color distribution uniformity of a single layer is judged, then the color distribution condition of the whole image is comprehensively analyzed to divide soil types, and different color recognition modes are selected according to the judged soil types to be compared with a colorimetric card digital image. Compared with a mode of directly extracting color information of an overall image and comparing a colorimetric card, hierarchical division is adopted, single-to-overall soil type judgment is carried out, and a proper recognition mode is selected according to a judgment result, so that the color recognition accuracy is effectively improved.
Owner:CHINA AGRI UNIV

Attention-based image generation neural networks

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for generating an output image. In one aspect, one of the methods includes generating the output image intensity value by intensity value according to a generation order of pixel-color channel pairs from the output image, comprising, for each particular generation order position in the generation order: generating a current output image representation of a current output image, processing the current output image representation using a decoder neural network to generate a probability distribution over possible intensity values for the pixel-color channel pair at the particular generation order position, wherein the decoder neural network includes one or more local masked self-attention sub-layers; and selecting an intensity value for the pixel-color channel pair at the particular generation order position using the probability distribution.
Owner:GOOGLE LLC

Systems and methods for reconstructing a three-dimensional object from an image

Embodiments described herein provide a 3D generation system from a single RGB image of an object by inferring the hidden 3D structure of objects based on 2D priors learnt by a generative model. Specifically, the 3D generation system may reconstruct the 3D structure of an object from an input of a single RGB image and optionally an associated depth estimate. For example, a radiance field is formulated to depict the input image in one viewpoint of the target 3D object, based on which other viewpoints of the 3D object can be inferred. Based on the visible surface depicted by the input image, points between the reference camera and the surface are assigned with zero density, and points on the surface are assigned with high density and color equal to the corresponding pixel in the input image.
Owner:SALESFORCE INC

Pattern wafer surface three-dimensional shape reconstruction method

The invention discloses a pattern wafer surface three-dimensional morphology reconstruction method, which comprises the following steps of: 1, constructing a small depth-of-field microscopic imaging system, and scanning the surface of a wafer along an optical axis direction to obtain a zoom image sequence; 2, establishing an image sequence space and a color vector space, calculating a pixel color gradient and a dispersion characteristic as a focusing evaluation basis, fitting a Gaussian distribution curve, obtaining a focusing evaluation curve with high steepness and an accurate center position according to a Bayesian prior distribution and posterior distribution principle, and performing focusing evaluation on the focusing evaluation curve; combining the curve peak value with the scanning interval to obtain a wafer surface initial depth map; and 3, performing filtering optimization on the initial depth map, and outputting a final three-dimensional reconstruction result. The method can effectively solve the problem that low-texture and low-contrast area definition evaluation is not accurate in a traditional focusing evaluation method, can improve the focusing peak searching speed and precision, and is suitable for high-precision microscopic three-dimensional reconstruction work of the wafer surface.
Owner:ANHUI UNIVERSITY OF ARCHITECTURE +1

Power equipment three-dimensional modeling and route planning method based on laser point cloud and image fusion

The invention provides a power equipment three-dimensional modeling and route planning method based on laser point cloud and image fusion. The method comprises the steps that an original point cloud sequence and an original image sequence collected by an unmanned aerial vehicle and pose information of the unmanned aerial vehicle are acquired; preprocessing the original point cloud sequence to generate a clean point cloud sequence; preprocessing the original image sequence to generate an enhanced image sequence; performing rough matching on the clean point cloud sequence according to the pose information to generate a panoramic scene point cloud; projecting points of the panoramic scene point cloud to the enhanced image sequence; mapping the pixel color to the panoramic scene point cloud by applying an L-M algorithm to generate a color point cloud; inputting the enhanced image sequence into a pre-trained SegFormer semantic segmentation model to obtain semantic probability distribution of each image; mapping the color point cloud to a corresponding image according to the semantic probability distribution to generate a semantic point cloud, the semantic point cloud including color information and semantic information; and generating a three-dimensional scene model according to the semantic point cloud to output an inspection route.
Owner:JIANGSU POWER TRANSMISSION & DISTRIBUTION CO LTD

Video compression platform based on AI

The invention relates to the technical field of video compression, in particular to an AI-based video compression platform, which comprises a key target detection module, an image region division module, a coding level setting module, a parameter adjustment and analysis module and a coding result integration module. According to the method, key region extraction is completed by collecting pixel color combination and contour change information in a video frame image, background and non-background region boundaries are delimited, a region attribute labeling result is constructed, a region distribution classification relation mapping coding level is established, and reference frame and prediction interval configuration is extracted. Real-time adjustment of compression levels is realized by performing normalized combination analysis on textures and motion change trends of different regions, and region fragments under different compression levels are recombined and subjected to unified code stream processing, so that coding resources can be dynamically allocated on the basis of accurately identifying contents in a compression process; the problems of inaccurate target positioning, fixed compression configuration, delayed content response and the like in an existing compression system are effectively solved.
Owner:HUNAN SANLI INFORMATION TECHNOLOGY CO LTD