139 results about "Texture enhancement" patented technology

The invention discloses a visual detection system and method for tiny flaws of industrial products, and belongs to the technical field of product detection, multi-source image data of a target industrial product under multiple detection angles and illumination conditions are acquired, and an image information matrix is established; performing region segmentation and texture enhancement on the image, and extracting local texture direction inconsistency parameters; carrying out normalization analysis on the pixel ratio under different spectrum channels, and calculating a multispectral reflectance ratio abnormal index; constructing a deep convolution recognition model; reasoning the image by using the model, and outputting a defect judgment result and a confidence score; judging whether the area is a flaw area based on a dynamic threshold mechanism, and outputting a detection report containing flaw position information and a visual heat map; according to the method, multi-dimensional fusion identification of texture structure disturbance and spectral response abnormity is realized, the micro defect identification precision is effectively improved, and the method has high robustness, automation and engineering practicability and is suitable for high-precision quality control requirements of various industrial scenes.

The invention discloses a two-way visual saliency detection method and device combining difference guidance and texture enhancement, and the method comprises the following steps: 1, obtaining an image to be subjected to saliency detection, and carrying out the marking and preprocessing of a data set; 2, constructing a visual saliency detection model which comprises a dual-path encoder (a saliency detection path and an image reconstruction path), an adaptive interaction network, a decoder network and an output network; a significance detection path in the dual-path encoder network uses a pre-trained ConvNeXt encoder, and an image reconstruction path uses VQ-VAE as a backbone network; the adaptive interactive network comprises a multi-scale convolution module and a gating fusion module; the decoder network comprises a mutual conversion attention module and a double-gating fusion module; the output network comprises a multi-level feature fusion module; 3, training the saliency detection model to obtain a trained saliency detection model; and 4, carrying out saliency detection on the image data by adopting the trained saliency detection model.

The invention discloses an image super-resolution method and system based on semantic perception tokens, and relates to the technical field of computer vision, and the method comprises the steps: generating semantic confidence and grouping information through the aggregation of content perception tokens, and decoupling a basic residual error into a texture enhancement and degradation inhibition guidance graph; in combination with a static semantic constraint mask and a sparse matrix multiplication mechanism, progressive focusing of attention is realized; a diffusion time step embedding and cooperative modulator is introduced, semantic guidance information is dynamically injected into a multi-step denoising process, adaptive attention features and diffusion reconstruction features are fused, and finally a high-fidelity and high-resolution image is output. According to the method, content-adaptive high-resolution image reconstruction is realized through collaborative modulation of a sparse attention mechanism guided by semantic grouping and diffusion denoising guided by semantic decoupling.

The invention discloses a multi-scale frequency-space fusion camouflage target detection method and system, and belongs to the technical field of camouflage target detection, an image and wavelet transform of the image are input into a WaveCamoNet model for camouflage target detection, and the model comprises a double-flow feature extraction module, a cross-domain dynamic fusion module, an edge texture enhancement module and a hierarchical decoder; inputting the image and the wavelet transform of the image into a double-flow feature extraction module, and extracting multi-scale spatial domain features and frequency domain features; inputting the multi-scale spatial domain features and the frequency domain features into a cross-domain dynamic fusion module, and performing multi-scale fusion and cross-domain dynamic fusion to obtain multi-scale fusion features and cross-domain dynamic fusion features; inputting the cross-domain dynamic fusion features into an edge texture enhancement module, and performing spatial calibration to obtain edge enhancement features; and inputting the multi-scale fusion features, the cross-domain dynamic fusion features and the edge enhancement features into a hierarchical decoder for decoding to obtain a binary detection image, and completing pixel-level positioning of the camouflage target.

The invention discloses a parallel U-Net-based dual-domain collaborative infrared and visible light image fusion method. The method comprises the following steps: firstly, extracting initial features of a source image by using dense connection blocks; then, parallel frequency domain branches and spatial domain branches are constructed, the frequency domain branches are combined with discrete wavelet transform and fast Fourier convolution to decompose and enhance multi-scale global frequency domain features, and the spatial domain branches capture long-distance spatial dependence with linear calculation complexity by using a convolutional layer and Mama based on a selective state space model; dynamic interaction and weighted fusion of double-domain information are realized through an adaptive feature fusion module; and finally, generating a fused image through an image reconstruction module. According to the method, the problems of high calculation overhead and video domain information negligence in the prior art are solved, and infrared heat radiation maintenance and visible light texture enhancement are effectively considered.

The invention relates to the technical field of image processing, in particular to an underwater crack segmentation-oriented color correction and texture sharpening double-branch enhancement system, which is characterized in that a data set is used for training, reasoning and analysis, and underwater crack images are stored in the data set; the system comprises an input unit used for receiving an underwater crack image; the color correction network is used for carrying out color correction on the underwater crack image and eliminating the problems of color deviation and low contrast of an underwater environment; the texture enhancement network is used for performing texture enhancement on the image after color correction; and the output unit is used for outputting the image enhanced by the texture enhancement network, and aims to recover the color and texture features of the underwater crack image through the color correction network and the texture enhancement network so as to enhance the perceptibility of the semantic segmentation model to the underwater crack.

The invention discloses a corn disease recognition method based on an improved generative adversarial network. The corn disease recognition method comprises the steps that S1, an initial low-resolution image of a corn field is collected and obtained through an unmanned aerial vehicle; s2, performing super-resolution reconstruction on the initial low-resolution image by using an improved generative adversarial network model; the model construction comprises the following steps: S2.1, constructing a shallow feature extraction layer; s2.2, constructing a deep feature extraction network based on a plurality of RRDB nested residual dense blocks; s2.3, constructing an attention module based on a space and channel dual attention mechanism; s2.4, a multi-scale texture enhancement module is constructed through multi-scale convolution and smooth branches; s2.5, constructing a global residual connection layer; s2.6, constructing an adaptive hybrid up-sampling module based on transposed convolution and stable up-sampling; s2.7, performing mapping output on the features after up-sampling; and S3, carrying out disease prediction on the high-resolution reconstructed image. According to the method, details such as spatial resolution and texture of the unmanned aerial vehicle high-altitude flight remote sensing image are improved, and then the corn disease monitoring precision is improved.

The invention relates to the technical field of image data processing, in particular to an image adaptive optimization processing method and system for laser printing output, and the method comprises the steps: carrying out the multi-scale feature analysis and fusion of an input original scanning image, and obtaining a fused feature distribution mapping matrix; performing adaptive contrast enhancement based on the fused feature distribution mapping matrix to obtain an enhanced contrast image; performing edge feature extraction on the enhanced contrast image by using an improved multi-direction edge detection algorithm to obtain an edge feature image with an enhanced edge; performing adaptive local texture analysis and enhancement on the edge feature image to obtain a texture enhanced image; and carrying out printing adaptability optimization based on the texture enhanced image to obtain a final optimized output image. According to the technical scheme, the quality of the image printed and output by the laser printing equipment is comprehensively and remarkably improved from multiple image processing dimensions.

The invention relates to the technical field of computer graphics and deep learning, in particular to a simulation bait automatic coloring method and system based on a 3D model. The method comprises the following steps: acquiring an uncolored 3D model and a reference image, and generating standard data through analysis verification, curvature grid division and image compliance detection; performing color conversion, texture enhancement and multi-scale downsampling on the compliant image to construct an image pyramid; performing multi-level feature extraction and adversarial training optimization based on a pre-trained convolutional neural network and a generative adversarial network, and generating an enhanced color texture map; performing UV expansion, color mapping and normal mapping fusion in combination with the model topology, and constructing an intermediate model with physical rendering attributes; and batch color consistency verification is realized through color histogram comparison, adaptive threshold segmentation and iteration parameter adjustment, and a standard model group is generated. According to the invention, efficient and highly realistic automatic coloring of the simulated bait is realized, and color consistency and rendering quality in batch production are guaranteed.

The invention provides a remote sensing image rotating target detection method based on dual-path feature enhancement, and relates to the technical field of computer vision and remote sensing image processing. The method comprises the following steps: constructing a dual-path feature enhanced remote sensing image rotating target detection network comprising a texture enhancement path and a direction modeling path; wherein in the texture enhancement path, a self-adaptive wavelet reconstruction module is adopted to enhance texture details and edge features in the input feature map; in the direction modeling path, performing spatial alignment and direction consistency modeling on the input feature map by adopting a multi-scale angle guide deformable encoder, and extracting features containing structure and direction information; fusing the features output by the two paths; constructing a joint loss function, and performing end-to-end training on the network; and detecting and positioning a rotating target in the remote sensing image by using the trained network. By adopting the method, the detection precision and robustness of multi-direction, multi-scale, densely distributed and small-size targets in the remote sensing image can be effectively improved.

The invention relates to the technical field of image enhancement, in particular to a dark field detail dynamic enhancement method and system for an LED backlight source, and the method comprises the following steps: setting a brightness threshold T1 and a brightness threshold T2 according to the brightness component of an input image frame, the dynamic backlight range of the associated LED backlight source and a gamma curve; according to the method, the pixel brightness is subjected to fine-grained division through partition judgment of the brightness component of the input image frame, so that more accurate dark field area positioning is realized, and the local linear transformation coefficient is constructed for the dark field mask area by using the original brightness component, so that the subsequent enhancement operation has continuity and edge retention characteristics; and adaptive classification of the image content is realized through a structure complexity threshold value, then differential enhancement strategies such as multi-direction texture enhancement, normal direction sharpening and curved surface smoothing are extracted respectively, and gain distribution conforming to different region characteristics is established, so that the dark field detail identification granularity is higher.

A method for lightweight image restoration is provided. The method includes receiving an input image of a scene captured at a pre-defined zoom level by an imaging sensor of the electronic device; inputting the input image into a naturalness restoration model to obtain a naturalness restored image and restored natural characteristics of the scene; inputting the input image into a texture enhancement model to obtain a texture enhanced image and enhanced texture characteristics of the scene; inputting the restored natural characteristics of the scene, the enhanced texture characteristics of the scene, and the input image into an image restoration model to obtain an intermediate enhanced image corresponding to the input image; and generating, using a fusing unit, an output image that is an enhanced version of the input image based on the intermediate enhanced image, the naturalness restored image, and the texture enhanced image.

The invention discloses a ventilation equipment blade quality detection method based on image processing, relates to the technical field of image processing quality detection, and is used for solving the problem of inaccurate blade defect detection. According to the method, the original image of the blade is obtained by adaptively adjusting exposure time, industrial camera gain and a controllable light source combination, reflection and shadow interference is inhibited by adopting extreme region masking, multi-frame compensation and local brightness balance, a contour enhancement image and a texture enhancement image are generated, an outer contour and a skeleton are extracted on the basis of a design contour, and the texture enhancement image is obtained. Re-sampling the texture along the skeleton and the normal direction, constructing a blade surface expansion view and a quality analysis coordinate system, detecting multiple types of surface defects in an expansion domain, forming a blade quality feature vector by combining chord length, elongation, section width and bending deviation, outputting a quality grade and a disposal suggestion, and realizing unified quantification of geometric deviation and surface defects of the blade. And the online detection stability and accuracy are improved.

The invention discloses a method for identifying early damage of a floor support plate. The method comprises the steps of image data acquisition and marking, texture enhancement preprocessing based on multi-scale gradient guidance, damage probability thermodynamic diagram generation fused with physical prior, double-flow feature interaction backbone network construction, and multi-scale deformable feature aggregation and damage perception channel weighted classifier design. And performing multi-task dynamic weighted loss optimization and model training. Image enhancement and mechanical strain prior information are fused in an input stage, and a self-adaptive interaction mechanism of global and local features is introduced into a network structure, so that the recognition capability of early damage such as microcracks, local corrosion and connection looseness is effectively improved. Experimental results show that the method shows higher detection precision and robustness under the conditions of complex backgrounds and multiple damage levels, and is remarkably superior to an existing convolutional neural network and a conventional detection model.

The invention relates to the technical field of lightweight rendering, in particular to a cloud-edge coordinated point cloud model lightweight rendering processing method and system. The method comprises the following steps: performing point cloud coordination prior processing according to electric power facility point cloud data to generate coordination prior point cloud data; a point cloud rendering model is constructed, point cloud curved surface coefficient calculation and curved surface reconstruction processing are carried out, and a point cloud curved surface reconstruction rendering model is generated; obtaining image view angle data, performing rendering model pose alignment processing and model texture enhancement mapping processing, and generating an enhanced texture rendering model; performing rendering coordination lightweight deployment and lightweight real-time rendering processing on the enhanced texture rendering model to generate lightweight real-time rendering data; and obtaining user behavior data, performing point cloud data rendering feedback analysis and cloud edge coordination caching in combination with the lightweight real-time rendering data, and feeding back to the terminal. According to the method, rendering of point cloud power equipment data curved surface reconstruction can be lightened, and redundant transmission of cloud edge rendering data is reduced and coordinated.

The invention discloses an automobile covering part surface abnormal defect synthesis and detection method based on illumination condition constraint. The method comprises the following steps: acquiring an automobile covering part defect data set; an illumination condition continuous mapping table is constructed, bilinear interpolation sampling is carried out on standardized sampling coordinates containing illumination conditions and defect types to obtain high-dimensional vectors, and encoding and mapping of defect sensing illumination conditions are achieved; a UNet adaptive modulation network integrated with an illumination perception residual block is constructed, abnormal texture enhancement is realized through the adaptive modulation network, and finally, the mask, the normal image and the abnormal texture enhanced image are fused to perform synthesis of an abnormal sample; a pre-trained backbone Wide ResNet50 is used as a feature extractor to carry out multi-scale feature aggregation on an illumination perception anomaly synthesis sample, a dichotomy discriminator is trained to discriminate the aggregated multi-scale features, and a trained model is used to carry out anomaly detection and positioning on a test chart. According to the method, the illumination condition is used as an optimizable control variable to be deeply fused into an abnormal synthesis and detection framework, illumination controllable abnormal synthesis, illumination robust feature learning and high-precision defect positioning are realized, and an effective solution is provided for industrial appearance detection under complex illumination.

The invention provides an end-to-end lightweight underwater target detection method and system, and belongs to the field of target detection, and the method comprises the steps: carrying out the frame sampling of an underwater video stream at a fixed interval, and obtaining an original image as a to-be-detected image; sending the original image into a differentiable brightness enhancement module and a lightweight texture enhancement module in parallel to generate a color correction image and a texture enhancement image in sequence, and inputting the original image, the color correction image and the texture enhancement image into a space correlation gating attention fusion module to obtain a fusion enhancement image; inputting the fusion enhancement graph into a backbone network formed by connecting a plurality of feature extraction layers in series and then connecting the feature extraction layers in series with a spatial pyramid pooling layer, and outputting a multi-scale feature graph; sending the multi-scale feature map into a path aggregation network to obtain a fused feature map with consistent semantics; and inputting the fusion feature map and the output of the spatial pyramid pooling layer into a multi-branch detection head, completing category, bounding box and confidence prediction, and outputting a final detection result through non-maximum suppression.

The embodiment of the invention provides a model training and image processing method and device, a storage medium and a program product. In the embodiment of the invention, the target degradation parameter is obtained by optimizing the initial degradation parameter with the target that the texture loss after image degradation processing is smaller than or equal to the set texture loss threshold value, so that the target degradation parameter is obtained according to the target degradation parameter obtained through optimization. The texture loss of the input image of the model obtained by carrying out degradation processing on the target image is smaller than the texture loss threshold compared with the target image, so that a moderately degraded training sample is generated, and the degraded image still keeps some texture information to provide reliable context clues for the model; and texture reasoning can be carried out based on real observation instead of generating some artifacts, so that texture enhancement is effectively realized.

The invention relates to the technical field of image processing and crack recognition, in particular to a cooling tower crack intelligent recognition system based on an unmanned aerial vehicle inspection image, which comprises an image acquisition module, a self-adaptive distortion correction module, a local texture enhancement module, a crack feature focusing module and a geometric morphology recognition and positioning module. Wherein the image acquisition module is used for acquiring an original image of the cooling tower; the self-adaptive distortion correction module is used for carrying out nonlinear geometric correction; the local texture enhancing module is used for enhancing the texture in the crack direction and inhibiting concrete background noise; the crack feature focusing module is used for extracting a pixel-level crack candidate region; and the geometrical morphology identification and positioning module outputs a crack data set with spatial position information. According to the method, through combination of image geometric correction, texture enhancement and three-dimensional space mapping, accurate extraction, spatial positioning and actual width measurement and calculation of the cooling tower cracks are realized, and the accuracy and reliability of structural defect identification are improved.

A method for lightweight image restoration is provided. The method includes receiving an input image of a scene captured at a pre-defined zoom level by an imaging sensor of the electronic device; inputting the input image into a naturalness restoration model to obtain a naturalness restored image and restored natural characteristics of the scene; inputting the input image into a texture enhancement model to obtain a texture enhanced image and enhanced texture characteristics of the scene; inputting the restored natural characteristics of the scene, the enhanced texture characteristics of the scene, and the input image into an image restoration model to obtain an intermediate enhanced image corresponding to the input image; and generating, using a fusion unit, an output image that is an enhanced version of the input image based on the intermediate enhanced image, the naturalness restored image, and the texture enhanced image.

Invention Name: A method and system for edge-enhanced remote sensing image segmentation that integrates attention and spatial state models Abstract: The present application discloses a method and system for edge-enhanced remote sensing image segmentation that integrates attention and spatial state models. The implementation steps are: constructing an edge texture feature enhancement structure; introducing the edge texture enhancement structure into the SegNext semantic segmentation model; dividing the remote sensing image segmentation dataset to generate a training sample set, a verification sample set, and a test sample set; preprocessing the dataset; using a neural network to preliminarily extract fine features of the optical remote sensing image, and then training the model with an edge texture enhancement decoder of the channel attention and spatial state model; finally, sending the test sample data to the edge texture enhancement model of the trained attention and spatial state model to obtain the test results. The patent of this invention utilizes the constructed edge texture feature enhancement module and the SegNext semantic segmentation model for collaborative training, which enhances the edge texture features while ensuring the features of the ground objects, thereby improving the accuracy of segmentation.

The invention provides an electronic license anti-counterfeiting storage verification method based on OCR and microscopic texture feature fusion. According to the method, OCR text semantic recognition and physical feature extraction are combined, and the unicity defect that a traditional pure OCR technology is prone to being attacked by content tampering and a pure Hash technology cannot recognize semantic tampering is effectively overcome by binding text content with non-replicable physical features such as paper fiber distribution and ink dot diffusion. In the feature extraction stage, a dynamic semantic alignment module is designed, high-value semantic areas such as identity card numbers and signatures can be synchronously marked in the OCR text positioning process, the texture Hash algorithm is guided to preferentially process the high-risk areas accordingly, the protection strength of key information is improved, the multi-scale texture enhanced Hash algorithm is proposed, and the accuracy of feature extraction is improved. By fusing the advantages of frequency domain wavelet energy analysis and spatial domain LBP feature extraction, even under the condition of low resolution caused by mobile phone shooting, anti-degradation extraction of texture features can still be realized, and the accuracy of counterfeit detection is improved.

The invention relates to the field of artificial intelligence, and discloses an artificial intelligence assisted neodymium-iron-boron magnet defect detection system which comprises the following steps: acquiring a magnet image and preprocessing process parameters; extracting texture enhancement features through multi-scale texture perception; adaptively generating or adjusting a defect detection model based on the process parameters; and carrying out defect classification positioning, discrimination and optimization. The system comprises an image acquisition module, a process parameter acquisition module, an image preprocessing module, a multi-scale texture perception feature extraction module, an adaptive model generation and selection module, a defect detection module, a discrimination output module and a model optimization module. Defects and background textures can be effectively distinguished, the false detection rate is greatly reduced, the robustness, universality, detection precision and stability of the system are improved, automation and intelligence are achieved, and efficiency and quality consistency are improved.

The invention discloses a space-time image texture angle detection and discrimination method and system, and belongs to the field of visual flow measurement instruments. According to the design, a Zynq series processor with an ARM + FPGA architecture is used as a core, and a system module comprises a PS host control end, a PL end texture enhancement module, a two-dimensional Fourier transform module, a frequency spectrum logarithm transform module, a Gaussian filter module and an image segmentation and texture angle detection and judgment module. Wherein the texture enhancement module inhibits noise interference and enhances the continuity of effective textures; the two-dimensional Fourier transform module converts the information of the spatial domain image to a frequency domain; the frequency spectrum logarithmic transformation module is used for carrying out logarithmic magnitude spectrum calculation on the transformed spectrogram so as to improve the visibility of a low-amplitude component; and the image segmentation and texture angle detection and discrimination module carries out angle detection and discrimination after effective and invalid signals are separated, and finally outputs an angle detection result. The method can be applied to a river surface flow velocity measurement system based on a space-time image velocity measurement method, and the adaptability and efficiency of flow measurement in complex scenes can be improved.

The invention discloses a plastic product flaw online detecting and sorting system based on machine vision, and particularly relates to the technical field of plastic flaw detection.The plastic product flaw online detecting and sorting system is characterized in that an image acquisition module acquires multi-angle original images under polarized light illumination and establishes a reflection response indexing relation; the spectral reflection structure shaping module implements differential illumination processing based on the indexing relation to generate a primary processing image; the feature enhancement processing module constructs three types of defect structure response channels, and generates a defect feature enhanced image through multi-scale texture enhancement; the defect identification module introduces a material adaptation factor and combines plastic physical characteristic correction to realize accurate defect discrimination; and the intelligent sorting execution module completes online grading sorting according to the defect identification indexes. The system is suitable for plastic products with different transparency, refraction coefficients and surface roughness, the problems that traditional detection is poor in adaptability, low in precision and inaccurate in sorting are effectively solved, the industrial real-time production requirement is met, and the product quality and the production efficiency are improved.

The invention discloses a casting riser image recognition method based on machine learning, and the method comprises the steps: carrying out the texture enhancement of a casting gray level image based on cross guide filtering, and forming a texture enhancement image data set; parallel straight line textures of the enhanced image are detected in a mode of combining a directional Gabor filter bank and direction consistency analysis; performing connected region analysis on a potential riser region, and performing triple screening through an area range, an aspect ratio range and a brightness contrast ratio; extracting an area feature, an edge feature, a texture feature and a shape feature of the riser candidate region, and marking the type of the riser candidate region to form a training sample set; a random forest classifier is introduced, the training sample set is input into the random forest classifier for training, a riser classifier is obtained, and the types of classified risers in the casting image are obtained; accurate positioning and classification of the casting risers are achieved, technical support is provided for automation and intelligentization of casting cleaning, and remarkable practical value is achieved.

The invention provides a depth estimation method based on a double-branch deep network and multi-attention fusion, relates to the technical field of image processing, is applied to a depth estimation network trained through self-supervised learning in advance, and comprises a detail branch module, a global branch module, a texture enhancement module, a sliding window self-attention module and a depth prediction head. The method comprises the following steps: acquiring an image needing depth estimation; respectively inputting the image into a detail branch module and a global branch module to obtain an initial detail feature and an initial global feature; inputting the initial detail features and the initial global features into a texture enhancement module to obtain enhanced detail features of the image; inputting the initial global feature into a sliding window self-attention module to obtain an enhanced global feature of the image; the enhanced detail features and the enhanced global features are spliced and then input to a depth prediction head to obtain a depth map, and the accuracy of depth estimation is effectively improved by fully extracting the detail features and the global features in the image.

The invention discloses a low-light image enhancement method based on multi-branch feature fusion. The method comprises the steps of obtaining low-light and normal-light images, and dividing training and test sets; processing the normal light image to obtain a single-channel color-enhanced, brightness-enhanced and texture-enhanced component graph; obtaining a multi-branch feature fusion low-light image enhancement neural network model; and inputting a to-be-measured low-light image into the model to enhance the backbone network, and outputting a restored image under a normal illumination condition. According to the invention, a color dynamic calibration module is provided, and a plug-and-play color correction normal form is provided for the model; an external source wound modulation module is provided, and key information can be focused while interference information such as noise is restrained. A brand new shared convolution is provided, so that the model can adaptively adjust the characteristic response intensity of each channel. An exogenous creative shared residual block structure is constructed, so that the enhanced backbone network can utilize the internal feature abstraction ability and the external auxiliary information guide advantage at the same time.

The invention relates to a generative adversarial network remote sensing image super-resolution reconstruction method based on replacement self-attention, and belongs to the technical field of image processing. The method comprises the following steps: constructing a PSA-ESRGAN generator network and an improved VGG discriminator network, inputting a high-resolution remote sensing image, performing down-sampling processing, inputting the processed low-resolution image for feature extraction, outputting a super-resolution image, performing multi-scale feature extraction by using the improved VGG network, and feeding back the output of a discriminator to the generator network. And the generator network improves the quality of the generated image by updating the weight of the convolution kernel. The method has the advantages that by combining the global feature enhancement capability of the PSA and the multi-scale feature retention mechanism of the improved VGG network, the detail recovery capability of the high-resolution remote sensing image is remarkably improved, meanwhile, the rationality of the overall structure is kept, the method is excellent in detail recovery, texture enhancement and edge quality improvement, and the method is suitable for popularization and application. And the robustness of the model and the adaptability to complex remote sensing images are improved.

The invention discloses an optical cement surface defect detection method and system based on visual inspection. The method comprises the steps that an original image set is acquired and preprocessed to obtain an intermediate image sequence; performing dynamic noise estimation on the intermediate image sequence, and selecting a high-frequency noise region in the intermediate image sequence for optimization to obtain a stable image sequence; defect detection is carried out on the stable image sequence, edge texture enhancement processing is carried out after preliminary defect area distribution is obtained, coordinates are mapped, and defect positioning coordinates are obtained; performing real defect judgment on the image according to the defect positioning coordinates to obtain a defect aggregation degree, dividing a high-density aggregation region, extracting the defect size of the high-density aggregation region, and generating a defect distribution mapping graph in combination with the defect aggregation degree; and carrying out image superposition according to the mapping graph and the original image set, and carrying out risk assessment on the high-density defect area according to the superposed image to obtain a final defect detection report. According to the method, accurate detection and analysis can be ensured when tiny defects are processed.