140 results about "Image separation" patented technology

The invention relates to a light-weight low-light image enhancement method based on illumination iterative adjustment, and the method comprises the steps: carrying out the HVI color space transformation of a to-be-processed image, obtaining an HVI image, separating an illumination intensity component, obtaining an illumination intensity feature map, and carrying out the convolution processing of the HVI image, and obtaining a shallow feature map; feature fusion and brightness adjustment are carried out on the illumination intensity feature map and the shallow layer feature map to obtain a depth feature map, up-sampling is carried out step by step through two layers of decoders to recover the size, channel adjustment is carried out on each layer of decoder by adopting convolution of a specified size, and up-sampling is realized through bilinear interpolation; a Mama-based structure refining module is introduced into the tail end of each layer of decoder to improve the structure restoration capability, the number of channels of HVI features output by the last layer of decoder is adjusted to a specified value through a convolution of a specified size, residual connection is carried out on the HVI features and an initial HVI image, an enhanced HVI image is obtained, and a restored image is obtained through inverse HVI transformation. Adaptive enhancement is carried out under various low light conditions, multi-directional spatial context aggregation is realized, and texture representation is improved.

The invention discloses an intelligent detection system for impurities and defects of raw coffee based on deep learning. The intelligent detection system comprises a visual detection module, a smell detection module, an optical detection module, a data management module, a data output system and a display device, the visual inspection is used for observing the surface of the raw coffee; the smell detection is used for detecting the taste of the raw coffee; and optical detection, which is used for analyzing and measuring the material composition and structure of the raw coffee and detecting whether foreign matters are contained in the coffee beans. According to the intelligent raw coffee impurity and defect detection system based on deep learning, the image processing module is arranged, images can be detected in all directions, the images are segmented through an image separation algorithm, the processing efficiency can be improved, meanwhile, the data management module can store and classify data, an independent database can be established, and the detection accuracy is improved. The coffee beans can be detected more accurately through internal data, and the detection effect of the device can be 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 provides a device for generating multiple beams of time sequence light based on a single-circle light spring, and relates to the technical field of ultrafast imaging, and the device comprises a leading edge tilt pulse generation system which is used for generating a long elliptical leading edge tilt pulse; the light spring generation system comprises a polarizer, a first wave plate, a diaphragm, a geometric conversion device, a convex lens, a phase compensator and a second wave plate which are sequentially arranged along the pulse propagation direction, and generates a single circle of light spring pulse based on the long elliptical leading edge tilt pulse output by the leading edge tilt pulse generation system; and the time sequence light beam generation system is used for dividing the single-circle light spring pulse into a plurality of beams of time sequence light through a circular porous diaphragm and a lens group. According to the invention, multiple beams of time sequence pulses can be generated as a coherent shutter to interfere with a signal light beam carrying ultrafast event information, multiple beams of time sequence light can be efficiently generated, multi-beam interference noise is optimized, and efficient image separation of ultrafast imaging is realized.

The invention relates to the technical field of image recognition, in particular to a checkerboard corner detection method and system based on image texture, and the method comprises the steps: obtaining a to-be-detected image; converting the image into a grayscale image; projecting a grid in the grayscale image to obtain a grid region; determining a gray jump point on each grid boundary in the grid region; and determining checkerboard angular points according to the gray jump points. According to the technical scheme, the checkerboard angular points are obtained by projecting the grid area and combining judgment of the gray jump points. Compared with the prior art, deep processing such as image separation of the image is avoided, so that the accuracy of the obtained checkerboard angular points is improved.

The invention discloses a method and a device for constructing a multilevel power field knowledge base, relates to the field of artificial intelligence, and mainly aims to solve the problem that the field knowledge base constructed based on an existing method is poor in effect of improving the application reliability of a large language model. Comprising the steps of performing text image separation on original data to obtain a domain knowledge original text and a domain knowledge original image; performing adaptive semantic text block segmentation on the original text to construct a semantic text block knowledge base; performing dual-channel collaborative knowledge extraction on the text blocks, fusing extraction results, and constructing a knowledge graph knowledge base; performing adaptive text description on the domain knowledge original image to generate a constructed image and a text description knowledge base; vector conversion and mapping alignment are carried out on the domain knowledge text blocks, the knowledge graph triple, the domain knowledge original image and a combination of corresponding text descriptions to construct a semantic vector knowledge base; and constructing a multi-level domain knowledge base based on the plurality of knowledge bases.

This application discloses a stroboscopic image processing method and apparatus, an electronic device, and a readable storage medium. The method includes: obtaining a first image photographed under a stroboscopic light source, where the first image is a raw RAW domain image; performing high-frequency image and low-frequency image separation on the first image, to obtain a first high-frequency image and a first low-frequency image; determining a target mask based on the first low-frequency image, where a mask value in the target mask is negatively correlated with brightness and banding strength of a region corresponding to the mask value in the first low-frequency image; filtering banding in the first low-frequency image based on the target mask, to obtain a second low-frequency image; and superposing the second low-frequency image and the first high-frequency image, to obtain an output image.

The invention relates to a wallpaper surface defect detection method, which comprises the following steps of: acquiring a wallpaper surface image by using a line-scan digital camera, correcting the image, establishing an average background model, and performing reflection compensation based on the average background model; s200, the image after reflection compensation is separated into R, G and B color channels, a gray level co-occurrence matrix (GLCM) of each channel is calculated to extract contrast features, the dynamic weight of each sub-block is calculated, a weighted RGB image is obtained, then the RGB image is converted into an LAB color space, an L channel is extracted, and # imgabs0 # is obtained; fusing the weighted RGB image and the L channel; carrying out three-layer decomposition by using a Daubechies8 wavelet basis, and carrying out reconstruction to obtain a background image; and carrying out dual-channel Ostu segmentation on the defect candidate image, carrying out decision fusion after segmenting the two channels, and carrying out noise removal and non-defect region filtering to obtain a defect region.

The invention relates to the technical field of computer vision, and discloses an image enhancement and optimization method and system based on multi-scale decomposition and a storage medium, and the method comprises the steps: decomposing a source infrared image into a basic layer and a detail layer, and obtaining an infrared image; dividing the source infrared image into a plurality of sub-regions for equalization, and obtaining a histogram equalization priori knowledge image; taking the source infrared image as a current input image, separating the current input image, storing the current input image into a Gaussian pyramid from bottom to top, and storing a residual image lost by sampling in each layer of the Gaussian pyramid into a Laplacian pyramid from bottom to top; extracting a down-sampling image from the top layer of the Gaussian pyramid, performing gamma transformation on each layer of residual image in the Laplacian pyramid to obtain an up-sampling enhanced image, and obtaining a fusion reconstruction priori knowledge image; and optimizing the infrared image according to the infrared image, the histogram equalization priori knowledge image and the fusion reconstruction priori knowledge image.

This invention discloses a laser image signal-to-noise ratio (SNR) enhancement method based on single-camera dual-channel imaging, relating to the fields of optical microscopy, digital image processing, and machine learning. The method includes: acquiring the original image from a single camera; separating the original image into a green channel image and a red channel image; constructing a denoising model based on the linear relationship between the green and red channel images, using the minimization of background noise in the green channel image as the objective loss function of the denoising model, and solving for an optimized parameter set; processing the green and red channel images according to the optimized parameter set to obtain a high SNR laser image after removing fluorescence background noise. This invention can remove fluorescence background noise from the original image acquired by a single camera without increasing hardware costs, effectively eliminating spectral crosstalk in the laser channel and significantly improving the laser image signal-to-noise ratio.

The application provides a low-light image enhancement method and system based on sequence multi-modal guidance, belonging to the technical field of image processing, wherein the method comprises: collecting a low-light RGB image and corresponding event data stream; encoding the event data stream into a three-dimensional event tensor; separating the low-light RGB image into low-frequency features and high-frequency features; inputting the high-frequency features and the three-dimensional event tensor into a pre-trained event feature extraction network to generate event features, and inputting the low-frequency features and the low-light RGB image into a pre-trained image feature extraction network to generate image features; and performing sequence enhancement on the image features and the event features to obtain an enhanced image. The application solves technical problems such as multi-modal feature fusion imbalance, texture fidelity decline caused by event data noise, and high model calculation complexity, and realizes high-quality enhancement of low-light images.

A method of processing a mammogram image, including receiving a digital mammography image of a breast of a patient, and separating the digital mammography image into three channels. The method includes removing noise and uneven illumination to form a first set of three uniform background images, and separating a luminance channel and a chrominance channel from the first set with a transfer function to form a second set of three uniform background images. The method includes applying principal component analysis to the second set; and assembling a single grayscale image and applying a first subspace projection. The method includes removing at least a portion of a pectoral muscle from the grayscale image, and normalizing the muscle removed image using a second order Gaussian Laplacian and an oriented diffusion filter to form a processed image. The method includes identifying cancerous and / or noncancerous regions in the processed image.

The application relates to the technical field of surface defect intelligent detection, and discloses a surface defect detection method and device based on multi-view synthetic imaging. The method comprises the following steps: acquiring a multi-view synthetic image without dead angles of a detected target after quality evaluation; performing highlight suppression and shadow enhancement on the multi-view synthetic image to obtain an enhanced multi-view synthetic image; separating the enhanced multi-view synthetic image into corresponding multiple single-view images, respectively performing defect detection on each single-view image to obtain corresponding single-view defect detection results, and generating a final overall detection result based on the single-view defect detection results. Through multi-view synthetic imaging, image quality control, high-reflective surface enhancement processing and multi-view joint defect detection, the application can effectively solve the problems of light interference and defect missing detection in high-reflective surface detection, and can realize high-precision surface defect detection with no dead angle coverage in a single detection.

This invention discloses an integrated broadband communication and stereo remote sensing method based on multi-satellite cooperation. For a ground target area, a central satellite and several neighboring auxiliary satellites are selected from a low-Earth orbit (LEO) satellite constellation. Each satellite collaboratively transmits radio frames towards the joint beam coverage area. Ground users receive the shared pilot signal embedded in the radio frames to obtain channel state information and recover data. Simultaneously, the central satellite receives and stores the ground echo signal corresponding to the shared signal and performs channel estimation. Combining the Doppler center differences between echoes from different cooperating satellites, multi-view range image separation is achieved. Furthermore, a stereo synthetic aperture radar (SAR) imaging method is used to calculate the three-dimensional geographic coordinates of reflectors in the target area. This method achieves integrated collaboration of communication and remote sensing functions within a LEO satellite constellation, contributing to improved constellation resource utilization efficiency and three-dimensional observation capabilities of the target area.

The invention provides a disparity map determination method, a three-dimensional reconstruction method and a related device. The disparity map determination method comprises the following steps: acquiring images of a to-be-detected object under projection of a plurality of images with different illumination intensities acquired by a camera module in the camera equipment, and obtaining a plurality of structured light object images of the camera module; the camera shooting equipment at least comprises two camera shooting modules; based on the plurality of structured light object images of the camera module, separating out a specular reflection image and / or a diffuse reflection image of the camera module; searching matching points based on the specular reflection images and / or diffuse reflection images of the two camera modules to obtain a first matching point pair; and based on the first matching point pair, obtaining a disparity map of the two camera modules. According to the invention, the quality of the disparity map on the highly reflective surface can be improved.

The application relates to the technical field of image processing, in particular to an intelligent detection method for aluminum alloy plate bending processing, which comprises the following steps: collecting a bending image sequence; separating each frame of image into a base image containing material texture information and a reflection image containing highlight information, to obtain a base image sequence; performing optical flow analysis and texture analysis on the base image sequence to identify processing defects generated in the bending process of the aluminum alloy plate; and generating a visual result image labeled with the positions of the processing defects and the positions of the highlights based on the processing defects and the reflection image, so as to realize detection of the bending defects of the aluminum alloy plate. According to the technical scheme, the light interference can be separated from the root cause, and finally, stable and reliable detection of the bending defects of the aluminum alloy plate is realized.

The invention relates to the technical field of PCB production detection, and discloses a method and a system for PCB multilayer board alignment deviation identification. According to the method, the multi-layer target theoretical projection image is generated through physical modeling, frequency domain analysis is carried out in combination with the X-ray detection image and the multi-layer target theoretical projection image, and quantitative evaluation of the reliability of different areas of the image is achieved through the obtained pixel-level credibility weight matrix. Further, through a method of deeply combining credibility evaluation and image separation, noise in the image is effectively suppressed, structural information of a high-credibility region is reserved, and accurate and reliable interlayer alignment deviation data can be provided for alignment deviation recognition through the obtained single-layer target separation image and target design parameters; the problem that the alignment recognition result is inaccurate due to aliasing distortion when an X-ray detection image is directly analyzed by a traditional method is solved, and the recognition accuracy of the alignment deviation condition of the PCB is remarkably improved.

The invention discloses a shooting signature processing method, computer equipment, a product and a storage medium, and the method comprises the steps: carrying out the preprocessing of a shot image in response to a triggering operation of the shot image containing the signature of a user, and obtaining a preprocessed shot image; performing shadow removal on the pre-processed shot image to obtain a shot image after shadow removal; performing edge detection on the shot image after shadow removal to obtain a candidate contour corresponding to the user signature and contour information of the candidate contour; and according to the candidate contour and the contour information of the candidate contour, cutting the shot image after shadow removal to obtain a signature image signed by the user. According to the technical scheme, the image definition and the user signature image separation precision are remarkably improved by processing the shot image containing the user signature.

In order to solve the technical problems that the existing droplet observation device cannot monitor the multi-nozzle printing process and the overall observation precision is low, the present application provides a multi-nozzle single droplet real-time observation device, an observation method and a 3D printing system.The observation device of the present application can make the observation camera capture the current working printing nozzle and automatically focus on it in real time through the cooperation of the first steering engine and the displacement platform, and complete the real-time monitoring of the printing state of the printing nozzle; through the high-frequency multiple flashing of the LED flash, the observation camera can capture the images of the same droplet at multiple times in one shot, at the same time, the high-frequency deflection of the piezoelectric deflection mirror and the telecentric lens are matched with each other, the overlapping images of the single droplet at adjacent moments are separated, so that a single photo can have enough droplet images to show the forming process, the overall observation precision is improved, the quality monitoring and fault diagnosis problems during additive manufacturing are improved, and the waste rate of electronic 3D printing is reduced.

The invention provides an image processing method and device, equipment and a medium, and relates to the technical field of artificial intelligence, in particular to the technical field of computer vision, deep learning and the like. The method comprises the steps of identifying a main body object corresponding to a target area on an original image in response to a detected input target operation for the original image; performing expansion operation on the main body object to obtain an adjacent area of the main body object; inputting a region identification graph for identifying the positions of the main body object and the adjacent region in the original image and the original image into an image matting model to obtain a foreground mask; separating the original image into a foreground image and a background image based on the foreground mask; performing blurring operation on the background image to obtain a blurred background image; combining the fuzzy background image and the foreground image into a preview image; and outputting the preview image.

The invention relates to the technical field of image processing, in particular to an intelligent detection method for aluminum alloy plate bending processing, and the method comprises the steps: collecting a bending image sequence; separating each frame of image into a substrate image containing material texture information and a reflection image containing highlight information to obtain a substrate image sequence; carrying out optical flow analysis and texture analysis on the substrate image sequence so as to identify processing defects generated in the bending process of the aluminum alloy plate; and based on the processing defect and the reflection image, generating a visual result graph marked with a processing defect position and a highlight position at the same time so as to realize detection of the bending defect of the aluminum alloy plate. By means of the technical scheme, illumination interference can be separated from the source, and finally stable and reliable detection of the bending defects of the aluminum alloy plate is achieved.

The present invention relates to the field of image processing, and in particular to an image separation and recognition system based on a computer vision model, comprising an image receiving module, an image recognition module, a model scheduling module and a recognition result output module; the image receiving module is used to receive an image and pre-process the image; the model scheduling comprises a model library, and the model scheduling module is used to schedule a visual model; the image recognition module is used to recognize the image and separate various parts in the image according to the image received by the image receiving module and the model scheduled by the model scheduling module; the recognition result output module is used for the recognition result of the image recognition module; the scheme is conducive to evaluating the reliability of different visual models by calculating the confidence indicators of different visual models, and is conducive to obtaining a more reliable recognition result by combining the recognition results of different visual models and fusing the images based on this.

The application discloses a multimodal medical image heterogeneous signal synchronous fusion method and device and system, comprising the following steps: S1, acquiring a DSA video stream and an IVUS video stream; S2, performing cross-modal medical image registration based on a space-time Transformer according to the DSA video stream and the IVUS video stream, and realizing medical image heterogeneous signal synchronous fusion. The technical scheme of the application solves the problems of image separation, inaccurate positioning, system delay and the like in the prior art, and improves the precision and safety of interventional surgery.

Systems and methods are provided for implementing a Siamese neural network using improved “sub” neural networks and loss function. For example, the system can detect a granular change in images using a Siamese Neural Network with Convolutional Autoencoders as the twin sub networks (e.g., Siamese AutoEncoder or “SAE”). In some examples, the loss function may be an adaptive loss function to the SAE network rather than a contrastive loss function, which can help enable smooth control of granularity of change detection across the images. In some examples, an image separation distance value may be calculated to determine the value of change between the image pairs. The image separation distance value may be determined using an Euclidean distance associated with a latent space of an encoder portion of the autoencoder of the neural networks.

The invention relates to a structure-semantic double-graph collaborative SAR image quality evaluation method, which comprises the following steps of separating an SAR image, and constructing a double-channel quality measurement set for independent modeling; indexes are obtained, correlation calculation is carried out, an index map is constructed, and each node corresponds to one index; on the basis of the index map, the dependency relationship between indexes is calculated from the structure dimension and the semantic dimension through self-supervised learning, the dependency relationship of the two dimensions is integrated, and an image level score based on the integration result is output; and constructing a measurement index shared with system analysis and a measurement index special for information retrieval, and carrying out quality measurement evaluation: after extracting a structure sub-graph, integrating an IR special index which is not contained in SAR prior into a global graph. The method is superior to a reference model in the aspects of recognition accuracy, recall rate, precision and efficiency, and a solution with high universality and good interpretability is provided for cross-modal image quality evaluation.

The present application relates to an image acquisition device and a method. The device comprises: a first projector, configured to project a projection pattern of a first wavelength to an object to be acquired, such that said object presents a first wavelength-based projection image; a second projector, configured to project light of a second wavelength to said object, such that said object presents a second wavelength-based mark point image; and a light splitting component, provided between said object and a preset camera module, and configured to separate the light of the first wavelength and the second wavelength, so as to separate the projection image from the mark point image. The preset camera module comprises: a first camera module and a second camera module, the first camera module being configured to acquire the projection image, and the second camera module being configured to acquire the mark point image. The present disclosure solves the technical problem in the prior art that when projection images and mark point images are acquired at the same time, the mark point images are contaminated by the projection images.

The imaging system includes a multi-distance camera having an imager and a lens. The lens includes a plurality of regions, and each region is physically different from each other region. The controller is in communication with the multi-range camera. The controller includes a processor and a memory. The memory stores instructions configured to cause the processor to process an image received from the imager by separating the image into a plurality of different region images, processing each different region image using a processing program corresponding to a region of the region image being processed, the region images are recombined into a single processed image, and the single processed image is provided to at least one vision system.

The invention discloses a single-camera partitioned optical microscopic imaging device which comprises a bottom plate, a support is mounted at the top of the rear side of the bottom plate, and a mounting plate is mounted on the front side of the support; the sample table is mounted on one side of the top of the bottom plate and is used for clamping a sample; the transmission illumination assembly is mounted on the other side of the top of the bottom plate and is used for providing uniform illumination for the sample; and the microscopic imaging assembly is mounted at the bottom of the front side of the mounting plate. According to the invention, through the arrangement of the image separation assembly, the intermediate image of the sample can be separated into microscopic images of a plurality of channels, and the microscopic images can be imaged on the target surface of the camera in real time after being transmitted through different paths, so that a single camera can carry out multi-channel imaging without considering whether a plurality of cameras are aligned or not; according to the optical microscopic imaging device, the problem of difference of time resolutions of a plurality of cameras does not exist, the optical path of the optical microscopic imaging device is relatively simple in the propagation process, some very rapid dynamic events can be synchronously captured, and the use effect is good.

In an embodiment of the present invention, a magnetic resonance water-fat image separation method, device, imaging system and storage medium are proposed. The method includes: within an echo cycle, obtaining a first echo group under a first readout gradient polarity, and obtaining a second echo group under a second readout gradient polarity; wherein the first readout gradient polarity and the second readout gradient polarity are opposite, and the echoes in the first echo group and the echoes in the second echo group are one-to-one symmetrical with respect to the echo center position of the echo cycle; obtaining a first echo image group based on the first echo group data obtained in each echo cycle, and obtaining a second echo image group based on the second echo group data obtained in each echo cycle; performing Dixon water-fat separation calculation using the first echo image group and the second echo image group to obtain a water image and a fat image. The technical solution in the embodiment of the present invention can improve acquisition efficiency and signal-to-noise ratio.

Methods and systems for setting up care areas (CAs) for inspection of a specimen are provided. One system includes a computer subsystem configured for acquiring an area of interest image for a specimen from images of the specimen generated by an imaging subsystem. The computer subsystem segments the area of interest image based on intensity of pixels in the image thereby separating the image into pixels corresponding to one or more specimen structures in the image having an intensity different than other pixels in the image. In addition, the computer subsystem detects edges of the specimen structure(s) in the image based on the pixels corresponding to the specimen structure(s) and determines characteristic(s) of one or more CAs in the images of the specimen based on the detected edges. The computer subsystem further stores information for the CA(s) for use in inspection of the specimen.