49 results about "Background reconstruction" patented technology

The invention relates to a fire point detection method, device and system based on a multiband infrared image and a storage medium, and relates to the technical field of image data processing. The method comprises the following steps: acquiring multiband infrared image data, and preprocessing the multiband infrared image data; performing feature classification extraction and feature fusion on fire point features of a fire point region and background features of a non-fire point region in the preprocessed multiband infrared image data by adopting a preset fire point detection model to obtain a target fire point segmentation probability graph; wherein the preset fire point detection model comprises an encoder, a fire point segmentation decoder, a background reconstruction decoder and a fusion module; and determining a fire point pixel detection result according to the target fire point segmentation probability graph and the segmentation threshold.

The invention relates to the technical field of automatic driving, in particular to a dynamic scene simulation generation method for automatic driving, which comprises the following steps: step 1, scene initialization and static background reconstruction: utilizing an improved 3D Gaussian splash technology, fusing LiDAR point cloud prior information, reconstructing a large-scale static road environment, and performing semantic decoupling on a scene; step 2, dynamic object and behavior injection: endowing each dynamic object with an initial motion track represented by a learnable Bezier curve, outputting an adjustment instruction for behaviors of other traffic participants by using a generative AI model according to the semantic information of the current scene and a self-vehicle planning track, and finely adjusting a control point of the Bezier curve; the interactive authenticity is improved, human driving behaviors are simulated through generative AI, physical reasonability is ensured through explicit tracks, the test efficiency is improved, and the long-tail scene generation efficiency is improved by 50% or above through conditional generation.

The invention discloses a cable sheath microcrack image identification method based on deep learning. The method comprises the following steps: acquiring a cable sheath image and executing image preprocessing operation; inputting to an improved MAE model, and generating a background reconstruction image and a crack reconstruction image; pixel-level residual fusion is carried out to generate a background shielding image; performing pixel-level fusion on the background shielding image and the preprocessed image to generate a background suppression image; micro-crack recognition operation is executed, and a preliminary crack response heat map set is output through image feature extraction and crack region judgment; executing a heat map accumulative analysis operation, and constructing a multi-scale accumulative heat map; judging a pseudo response risk area according to the local response change rate; response value retraction operation is executed based on the corresponding local area, and a crack heat map after pseudo response suppression is generated; and extracting a high-confidence crack region to obtain a cable sheath microcrack identification result. According to the invention, the precision and robustness of microcrack detection are improved, and the background interference and false detection risk are reduced.

According to the chemical material detection method and system based on deep learning provided by the invention, the online adaptive training generative adversarial network BR-GAN is introduced, and the generator can dynamically learn background optical feature mapping under the fine tuning condition of production process parameters; meanwhile, a multi-layer convolution discriminator is used for recognizing time correlation of background optical features, so that a defect-free background reconstruction image matched with the current production condition can be generated, and a difference image is generated by calculating pixel-level residual errors and feature-level attention differences of a real-time image and the reconstruction background; microscopic bubble signals are effectively separated from complex backgrounds changing along with technological parameters, and the effect of accurately detecting microscopic bubbles is achieved.

The invention provides a robot multi-modal data enhancement system for an industrial close-range grabbing scene, and relates to the technical field of industrial robots. The system comprises a data preprocessing module which obtains an RGB image composed of a robot execution track and visual information, a depth image and track data; the background reconstruction module is used for carrying out background reconstruction on the RGB image; the adaptive illumination enhancement module generates illumination disturbance images of multiple versions for the RGB image after background reconstruction, and performs fusion processing to obtain an enhanced RGB image; the depth information reconstruction module is used for reconstructing and complementing missing or invalid areas existing in the original depth image; the multi-modal data synchronization verification module is used for carrying out consistency verification; and the enhanced data output module is used for re-packaging and outputting the enhanced RGB image, the depth image, the track data and the corresponding check code. According to the system, on the premise that existing hardware does not need to be replaced, the quality and robustness of multi-modal data are remarkably improved.

The application discloses a three-dimensional static background reconstruction method, device and equipment and a storage medium. Real images and camera parameters of a scene are acquired; initialized 3D Gaussian parameters and the camera parameters are input into a 3D GS renderer; a rendered image output by the 3D GS renderer and the real images are input into a mask predictor; the 3D Gaussian parameters and the mask predictor are trained; the trained 3D Gaussian parameters and the camera parameters are input into the 3D GS renderer, and a rendered image of a reconstructed static background is output. The application is based on the 3D GS renderer and the mask predictor, uses 3D GS to reconstruct a static background from scene data containing dynamic objects, and does not depend on artificial annotation information.

The application relates to an abnormality detection method based on a hyperspectral image. The main body is based on a space-spectrum combined feature extraction method of filtering and low-rank decomposition to perform abnormality detection on the hyperspectral image. The specific method comprises the following steps: firstly, in the spatial dimension, a reduced dimension image is obtained through a data dimension reduction and eigenvalue weighted fusion method, and then an improved spatial filtering method is used to extract the spatial features of the image to obtain an initial spatial feature image. In the spectral dimension, a background reconstruction image of the approximate background is obtained by using a Tucker decomposition method on the original hyperspectral image, and a background dictionary of the image is obtained by using an improved k-means clustering method, then the background dictionary is input into a low-rank decomposition model to obtain a sparse matrix, and an initial spectral feature image is obtained, finally, the initial spectral feature image is fused with the spatial feature image to realize abnormality detection.

The invention provides a model training method and device of a generative model, a generation method and device of the generative model, and equipment. The method comprises the steps of obtaining a target training sample; the target training sample comprises an initial two-dimensional image and mask data used for representing each initial object in the initial two-dimensional image; inputting the initial two-dimensional image and the mask data of each initial object into a to-be-trained generative model to obtain model features of a 3D mesh model representing the initial object in the initial two-dimensional image by using a mid-scene reconstruction branch of the to-be-trained generative model, and reconstructing a background reconstruction branch of the to-be-trained generative model to obtain a model feature of a 3D mesh model representing the initial object in the initial two-dimensional image; background features capable of describing a background area in the initial two-dimensional image are obtained; and performing model training on the to-be-trained generative model based on model features of a 3D mesh model representing an initial object in the initial two-dimensional image and background features describing a background region in the initial two-dimensional image to obtain a target generative model.

The application discloses a hyperspectral remote sensing image target detection method based on target-background reconstruction deviation, comprising the following steps: acquiring a to-be-detected hyperspectral remote sensing image, inputting the pre-trained asymmetric auto-encoding network, and obtaining the first output and the second output of each pixel in the image; the pre-trained asymmetric auto-encoding network comprises a feature extraction subnetwork, a feature fusion subnetwork and a feature reconstruction subnetwork; the first output is the output of the feature extraction subnetwork, and the second output is the output of the feature reconstruction subnetwork; the pre-trained asymmetric auto-encoding network is obtained by training a mixed target spectrum and a mixed background spectrum; the mixed target spectrum and the mixed background spectrum are generated based on a first hyperspectral remote sensing sample image, a prior spectrum of a preset target and a bilinear spectrum mixing model; according to the first output and the second output, the spectral angle distance of the pixel is determined; and the spectral angle distance is smoothed to obtain the detection result of the to-be-detected hyperspectral remote sensing image.

The invention belongs to the technical field of underwater sonar imaging and intelligent identification, and relates to a virtual sonar image generation method and system. The method comprises the following steps: mapping a real optical target image into a virtual sonar target image by adopting a vision-text fusion network; embedding the virtual sonar target image into the real sonar background image to generate a combined image; and background compensation and texture consistency adjustment are carried out on the combined image by using a virtual sonar image background reconstruction network to obtain a virtual sonar image with a background. According to the invention, the bottleneck that real sonar data acquisition is limited by environment, cost and equipment conditions is broken through; meanwhile, the target type and number can be flexibly expanded, and the problems that sonar samples are insufficient in category and unbalanced in distribution are effectively solved.

The invention discloses an intelligent switching method and system for face component structured mask learning, and the method comprises the steps: extracting a portrait mask with an edge perception capability, and reconstructing a high-fidelity background region with spatial continuity through combining a background reconstruction network; generating a standardized face image with geometric consistency by adopting a face detection and posture correction network; generating an optical head portrait with vision and texture consistency; performing structured mask learning on the standardized face image and the optical head image by using a semantic segmentation network to realize pixel-level face analysis; obtaining a multi-component region mask through mask calculation based on morphological operation; generating an optical head portrait with a natural transition effect by using an optical head portrait synthesis module based on Poisson fusion; and generating a hair changing portrait with physical authenticity by using a hair changing portrait synthesis module. According to the method, the visual fidelity, the detail consistency and the identity retentivity of the face image after hair changing are remarkably improved.

The invention discloses an event camera video reconstruction method and system based on active aperture modulation, and the method comprises the steps: introducing an aperture modulation strategy for the first time, and reconstructing an initial frame with good quality by periodically adjusting the opening and closing of an aperture and actively triggering a dense global event signal; the problem of low background reconstruction quality caused by sparse events in a static region in the prior art is solved, and good prior information is provided for subsequent dynamic scene reconstruction. By constructing a forward-reverse bidirectional network, rich intensity information is provided for a static scene, and the problems of background disappearance and error accumulation in long-time operation in a traditional method are effectively solved. And meanwhile, high-time-resolution capture of a dynamic region is kept, and high-fidelity and high-dynamic-range video reconstruction is realized.

The invention discloses a foreign matter segmentation method based on dynamic background suppression and weak supervised learning, and the method comprises the steps: 1) training a static background reconstruction network, screening a foreign matter-free frame from a video as the input of a full-connection automatic encoder, and learning a static background of a scene through minimizing reconstruction loss; 2) inputting a video frame to reconstruct a static background, performing adaptive background modeling by adopting a Gaussian model so as to perform foreign matter judgment, and applying a neighborhood consistency constraint to improve robustness; 3) identifying an area containing foreign matters and a dynamic background, and screening a sequence only containing the dynamic background to train a U-Net network, so that the U-Net network can predict a dynamic background mode; 4) performing pixel-by-pixel operation on the foreground binary image and the dynamic background probability predicted by the U-Net network to generate a foreign matter probability graph; and 5) performing morphological post-processing on the foreign matter probability graph to optimize the integrity of the foreign matter contour. According to the method, static and dynamic background features are learned through dual-network collaborative learning, and the foreign matter segmentation performance in a complex environment is remarkably improved.

The application discloses a kind of cable sheath micro crack image recognition methods based on deep learning, including the following steps: cable sheath image is collected and image pre-processing operation is executed;Input to improved MAE model, generate background reconstruction map and crack reconstruction map;Pixel-level residual fusion is carried out, and background mask map is generated;Pixel-level fusion is carried out to background mask map and pre-processing image, and background suppression map is generated;Micro crack identification operation is executed, and preliminary crack response heat map set is output by image feature extraction and crack area discrimination;Heat map cumulative analysis operation is executed, and multi-scale cumulative heat map is constructed;According to local response change rate, judge false response risk area;Response value retraction operation is carried out based on corresponding local area, and crack heat map after false response suppression is generated;High confidence crack area is extracted, and cable sheath micro crack recognition result is obtained.The application improves the precision and robustness of micro crack detection, reduces background interference and false detection risk.

This invention relates to a texture primitive extraction method for tire X-ray images, addressing the challenge of accurately obtaining pixel-by-pixel texture information at the individual cord level while suppressing background interference. It falls under the field of computer vision and image processing technology. The method combines frequency domain analysis to obtain texture direction and spacing, utilizes this information to construct a mesh mask, and further combines background point extraction with real background reconstruction to filter out specific background regions in the original image. Under directional constraints, the remaining texture mesh is continuously tracked to obtain pixel-by-pixel texture information, thus achieving texture primitive extraction. This provides a more reliable foundation for subsequent pathological detection, structural analysis, and cord-level texture modeling.

The invention provides a ground coffee foreign matter detection method and device based on background reconstruction, a medium and a product, and the method comprises the following steps: obtaining an original transmission image of to-be-detected ground coffee, and carrying out gray normalization to obtain a standard input image; respectively inputting a standard input image into a background reconstruction branch and a target detection branch; in the background reconstruction branch, generating a pure background prediction image by using a self-encoding model, and calculating a gray difference between the pure background prediction image and the standard input image to obtain a foreign matter residual image; convolutional features are extracted from the foreign matter residual image, and a spatial residual attention weight map is generated; in the target detection branch, performing multi-layer convolution on the standard input image, and performing weighted fusion in combination with the attention weight map to obtain a foreign matter sensitive feature map; and outputting a foreign matter detection result through prediction head network classification and regression decoding. By implementing the technical scheme provided by the invention, the detection sensitivity of low-contrast and low-density foreign matters is improved, and meanwhile, the false detection rate caused by complex powder textures is reduced.

The invention provides a dazzle light removing method. The method comprises the steps that a night light emitting model containing a dominant light source and a non-dominant light source is established through radiation transmission analysis; the method comprises the following steps: acquiring HDR night scene dazzle light damaged image data, and preprocessing and normalizing the data to obtain standardized data; aiming at input data, an unsupervised exposure guide label generation method is adopted, and a dazzle light area around a dominant light source and a non-dominant light source is positioned based on modulus imaging constraint; introducing a Uform network guided by an exposure label, combining with an exposure guide mask attention mechanism, and jointly optimizing a dazzle light and background decomposition and reconstruction process through exposure guide loss; based on a pseudo-label guided light source recovery method, decoupling a dominant overlapped light source under the constraint of exposure and connectivity, and retaining a non-dominant light source; and fusing the non-glare background reconstruction result and the light source reconstruction result to obtain output image data which is free of glare and contains an ideal light source. The method has the advantages of breaking through light source and dazzle light marking dependence, improving the reconstruction precision of the overlapped light source and the weak light source, enhancing the supervised learning ability and effectively removing night dazzle light artifacts.

The application relates to the technical field of computer vision, in particular to a NeRF map construction method and device based on visual SLAM, wherein the method comprises the following steps: in a dynamic scene, acquiring a static target feature point of a current frame image of a camera; tracking the static target feature point, solving a current frame camera pose of the camera, and selecting a key frame according to the number of projection inliers in a tracking process and tracking time; sampling pixel points in the key frame according to ORB features, inputting a key frame sampling result into a pre-established NeRF map model for training, obtaining NeRF map model parameters, and obtaining a NeRF map construction result. The application can generate a dense NeRF map in real time for a dynamic scene, overcome the problem of false matching in the dynamic scene, maintain simplicity with high efficiency, eliminate the influence of dynamic targets, complete high-quality background reconstruction, and effectively meet the positioning and mapping application requirements in different scenes.

This invention provides a multimodal data augmentation system for robots in industrial near-field grasping scenarios, relating to the field of industrial robot technology. The system includes: a data preprocessing module, which acquires an RGB image, a depth image, and trajectory data composed of robot execution trajectory and visual information; a background reconstruction module, which reconstructs the background of the RGB image; an adaptive lighting enhancement module, which generates multiple versions of lighting perturbation images from the reconstructed RGB image and fuses them to obtain an enhanced RGB image; a depth information reconstruction module, which reconstructs and completes missing or invalid regions in the original depth image; a multimodal data synchronization verification module, which performs consistency verification; and an enhanced data output module, which repackages and outputs the enhanced RGB image, depth image, trajectory data, and corresponding checksums. This system significantly improves the quality and robustness of multimodal data without requiring replacement of existing hardware.

The present application relates to a kind of hyperspectral image anomaly detection method, the present application is based on the background reconstruction subtraction of multiple feature joint to the hyperspectral image is carried out anomaly detection.Specific method includes: first using spatial feature and spectral feature joint improved double window three-edge filtering method to reconstruct background to image;Second, for the distribution characteristics of the abnormal target in the analysis image, on the basis of traditional saliency detection method, propose the saliency feature extraction method based on global context perception to extract the saliency feature map of image;Then the saliency feature map of image and the square difference of three-edge filtering after reconstructing background map are obtained Abnormal target initial detection map;Finally, using spectral correlation coefficient to obtain the spectral weight map of image, and it is fused with initial anomaly detection map to obtain the final anomaly detection result.

The application belongs to the technical field of underwater sonar imaging and intelligent identification, and relates to a virtual sonar image generation method and system. The method comprises: mapping a real optical target image to a virtual sonar target image by using a vision-text fusion network; embedding the virtual sonar target image into a real sonar background image to generate a combined image; and performing background compensation and texture consistency adjustment on the combined image by using a virtual sonar image background reconstruction network to obtain a virtual sonar image with a background. The application breaks through the bottleneck of real sonar data collection being limited by environment, cost and equipment conditions; at the same time, the target type and quantity can be flexibly expanded, effectively solving the problems of insufficient sonar sample categories and unbalanced distribution.

This invention relates to a model construction method for background reconstruction of Mura defect images on OLED screens, belonging to the field of Mura defect technology. It solves the problems of poor reliability in identifying Mura defects and inaccurate background reconstruction due to the extremely limited number of Mura defect samples during model training in existing technologies. The method includes: acquiring a defect-free image of the OLED screen; converting the defect-free image to grayscale to obtain a grayscale image; generating a defect-free grayscale image and a defect image based on the grayscale image; constructing a sample image set based on the defect-free grayscale image and the corresponding defect image; constructing a background reconstruction adversarial network model, which includes a background reconstruction network model and a discriminant network model; training the background reconstruction adversarial network model based on the sample image set to obtain a trained background reconstruction adversarial network model; and using the trained background reconstruction network model as the model for background reconstruction of Mura defect images.

This invention discloses a multispectral anomaly detection method and system based on a physically constrained spatial-spectral decoupling GAN, belonging to the field of spectral detection technology. The method includes: acquiring and preprocessing a multispectral image to obtain a standardized tensor block; inputting this block into a dual-stream generator to output an ideal background reconstruction image; inputting this reconstruction image and a real background image into a multi-scale block discriminator for adversarial discrimination, and sending the discrimination result into a model training and optimization module; calculating the gradient using a composite loss function integrating physical constraints, synchronously updating the parameters of the dual-stream generator and the multi-scale block discriminator to obtain a multi-scale dual-stream GAN model; inputting the multispectral image to be tested into the trained model to obtain the ideal background reconstruction image, calculating the anomaly score map between the reconstruction image and the original image, and outputting the anomaly detection binarized result after threshold segmentation. This application solves the problem of difficulty in simultaneously considering texture and spectral features in multispectral data, significantly improves the model's sensitivity to anomalous targets, and reduces the false alarm rate.

The embodiment of the present application provides a kind of modeling method, system, equipment and storage medium for extracting foreground object, belong to three-dimensional reconstruction technical field.Modeling method for extracting foreground object includes: obtaining the image point cloud data of target object and camera pose information;The foreground and background in image point cloud data are separated, and the foreground object point cloud is obtained;According to the foreground object point cloud extracted, the minimum bounding box of foreground is determined;Based on the above data, foreground point data and background point data are determined;Based on foreground point data and background point data, the preset reconstruction network is used to determine foreground object reconstruction data and background reconstruction data;According to foreground object reconstruction data and background reconstruction data, generate target object model and display.The embodiment of the present application calibrates foreground and background using minimum bounding box, reduces the cost of manual annotation, also uses preset reconstruction network to predict model data, improves the accuracy and accuracy of model reconstruction.

The invention discloses a solar cell surface defect intelligent detection system based on deep learning. The system comprises a data acquisition module; according to the method, detection design based on background reconstruction is adopted, the conditional depth reconstruction network is constructed through the original image, easy to obtain, of the surface of the defect-free solar cell, and the defect-free background reconstruction image can be reconstructed through the conditional depth reconstruction network; in the detection stage, the to-be-detected image and the residual image corresponding to the defect-free background reconstruction image can remarkably highlight a defect area which destroys the background regularity, and the defect can be positioned only through the original image of the surface of the defect-free solar cell. According to the method, the technical thought that the background is reconstructed first and then the residual error is calculated is adopted, the defect detection problem is converted from mass pixel-level defect labeling to normal background modeling, and the problems that labeling samples are scarce and labeling cost is high in an industrial scene are greatly relieved.

The invention discloses a hyperspectral anomaly detection method for screening multi-level spectral features based on self-learning, and the method comprises the steps: dynamically selecting a most representative wave band according to the statistical characteristics and structure of hyperspectral data through a self-learning wave band screening module; the hierarchical feature fusion module is used for effectively capturing spectrum differences and accurately bridging feature information of different hierarchies. And network model training is constrained through structural similarity loss and separation suppression loss. Interference of abnormal information in detection is effectively suppressed, the accuracy of background reconstruction of hyperspectral image anomaly detection is ensured, the separability of an abnormal target under a complex background can be effectively improved through the extracted discriminative features of the abnormal target, the false detection rate and the omission rate are reduced, the reconstruction precision of hyperspectral image anomaly detection is remarkably improved, and the detection accuracy of hyperspectral image anomaly detection is improved. And the accuracy and robustness of hyperspectral image anomaly detection are improved.

The application provides a wafer defect detection background suppression and defect feature enhancement method. In the face of serious background interference such as periodic texture and uniform noise in the wafer defect detection image, the method first uses frequency domain background reconstruction and removal technology to obtain a preliminary background removed image. Further, based on the pixel-level labeled defect data set, a self-encoder network architecture with a fusion spatial attention mechanism is designed to realize accurate positioning and feature weighting enhancement of the defect area, while suppressing the residual background noise. On this basis, a background-defect dual-region dynamic adjustment strategy is constructed to complete the adaptive balance fine tuning of the enhancement parameters, and then the cooperative optimization of wafer defect detection image background suppression and defect enhancement is achieved. Through the scheme, high-quality image preprocessing support can be provided for the wafer defect high-precision detection task, and the precision and efficiency of wafer defect detection are effectively improved.

The application discloses a kind of fabric hyperspectral flaw reconstruction method and system based on frequency domain dynamic convolution, belong to computer vision and hyperspectral imaging technical field.The method includes: input fiber end-member spectral library data to carry out physical constraint modeling;Frequency band decoupling convolution kernel is generated based on frequency domain dynamic convolution, linear mixing term and non-linear interaction term are combined to reconstruct background;Through the decomposition of high-frequency sub-band by learnable wavelet base, the flaw feature is enhanced using frequency domain dynamic convolution;Multi-scale feature fusion is realized by adaptive low-pass / high-pass filtering and bidirectional attention gate;Adversarial optimization is carried out based on generative adversarial network, and the flaw mask is output.The system includes: end-member spectral library module, background reconstruction module, flaw residual branch module, frequency domain perception multi-scale fusion module and adversarial optimization module.The application has the advantages of strong physical interpretability and high reconstruction accuracy, and is suitable for camouflage fabric hyperspectral flaw detection.

This application discloses a hyperspectral anomaly detection method and system based on low-rank constrained autoencoders. The method applies low-rank constraints to the hidden layer representation of an autoencoder network, mines high-order low-rank characteristics for background reconstruction, and uses reconstruction errors to identify anomalies. The method implementation steps include: Step 1: Inputting the hyperspectral data X to be detected, setting network parameters W and b, learning rate α, and initial values ​​for adjustment parameters λ1 and λ2; Step 2: Constructing a loss function for the hyperspectral data X using low-rank constraints; Step 3: Learning the network using the loss function, updating the network parameters {W, b}, and obtaining the reconstructed data of the input data X; Step 4: Constructing an error matrix E; Step 5: Detecting the reconstruction error matrix.

The application discloses a hyperspectral image anomaly detection method, device and equipment and a medium, relates to the field of hyperspectral anomaly detection, and comprises the following steps: inputting hyperspectral pretreatment data into a pre-constructed convolutional autoencoder with cross-connection layers, and outputting a background reconstruction tensor of the hyperspectral pretreatment data; stretching the background reconstruction tensor into a matrix along a spectral dimension, and constructing a target function for the matrix to optimize the network output of the convolutional autoencoder; iteratively updating the low-rank constraint and the sparse constraint of the target function by using an alternating direction multiplier method, so as to adjust the network parameters of the convolutional autoencoder; when the number of iterations reaches a convergence threshold, determining a feature tensor output by the current convolutional autoencoder; subtracting the feature tensor from an original feature tensor to obtain an error tensor after background reconstruction; square-summing and then taking the square root of the error tensor of each spectral dimension slice to obtain a detection map; and performing anomaly detection on the detection map to obtain an anomaly detection result.