51results about How to "Improve detail performance" patented technology

The invention discloses a corrugated pipe inner wall multi-type flaw detection method based on deep learning, and the method comprises the following steps: S1, collecting an initial corrugated pipe inner wall image and corrugated pipe inner wall images in different illumination environments, and generating an original image; s2, preprocessing the original image to generate a preprocessed image; s3, extracting surface geometric overall features of the inner wall of the corrugated pipe from the preprocessed image through a pre-trained ResNet101 network; s4, inputting the preprocessed image and the surface geometric overall features into an improved StyleGAN3 network to generate an enhanced image; s5, inputting the enhanced image into the multi-scale adaptive deep network, and outputting a defect detection result; and S6, performing post-processing on the flaw detection result, and outputting a final flaw detection result. According to the method, deep learning and geometric characteristic embedding technologies are adopted, so that the defects of the inner wall of the corrugated pipe are accurately detected, and the method has the advantages of high precision, high robustness and comprehensive adaptability.

The invention relates to the technical field of data processing and analysis, in particular to a three-dimensional model data management system and method for engineering surveying and mapping. The method comprises the following steps: acquiring engineering surveying and mapping data; preprocessing the engineering surveying and mapping data to obtain standard engineering surveying and mapping data; performing multi-dimensional feature recognition on the standard engineering surveying and mapping data to obtain multi-dimensional surveying and mapping feature data; performing key engineering surveying and mapping information extraction on the multi-dimensional surveying and mapping characteristic data to obtain key engineering surveying and mapping information; performing association analysis on the multi-dimensional surveying and mapping characteristic data and the key engineering surveying and mapping information to generate multi-dimensional engineering surveying and mapping associated data; monitoring the engineering surveying and mapping data in real time to obtain engineering surveying and mapping real-time monitoring data; through a data processing technology, a geographic information system technology and a three-dimensional mapping technology, integrated management of multi-source heterogeneous data is realized, and dynamic updating management of the data is realized, so that the presentation accuracy of the engineering surveying and mapping three-dimensional model is improved.

The embodiment of the invention discloses a three-dimensional reconstruction method based on structured three-dimensional Gaussian representation. The method comprises the following steps: acquiring a target image containing a target object at multiple visual angles; generating an unstructured Gaussian primitive representation corresponding to the target object according to the target image; establishing a transmission matrix describing the unstructured Gaussian primitive distribution and the structured Gaussian primitive distribution, solving the transmission matrix by using an optimal transmission algorithm, and generating a structured Gaussian primitive representation corresponding to the target object according to the optimal transmission matrix; obtaining a text description for the target object, and complementing the structured Gaussian primitives according to the text description by using a diffusion model obtained by pre-training; and performing three-dimensional reconstruction by using the complete structured Gaussian primitive. According to the method, the optimal transmission algorithm is used for carrying out structured representation on the three-dimensional Gaussian primitives, and text description is used for complementing the Gaussian primitives through the diffusion model, so that the quality of three-dimensional reconstruction is improved.

The invention discloses an image restoration method based on a multi-scale residual module and feature fusion, and belongs to the technical field of image processing. The multi-scale residual module designed by the invention can effectively capture feature information of different scales of the image, the quality of the repaired image is improved from a fine texture to a macroscopic structure, a feature fusion strategy fusing detail branches and semantic branches can fully integrate features of different levels, the detail expression and overall coordination of the image are enhanced, and the image quality is improved. Through fusion of multi-scale features, limitation possibly brought by single-scale features is avoided, and the repaired image is more real and fine in color, texture and structure. And the multi-scale residual error module can relieve the gradient disappearance problem, the calculation efficiency is high, the image restoration task can be completed in a short time, and the introduced and designed dual-channel global attention module has good robustness, so that the method can stably play a role in various complex image damage conditions.

The invention relates to the field of data processing, in particular to an environment perception data processing method and system based on an intelligent helmet. The method comprises the following steps: acquiring a scene image in front of a user and multi-source environment real-time sensing data; performing scene attribute state analysis on the multi-source environment real-time sensing data, and constructing a real-time scene attribute state diagram; performing contrast dynamic balance adjustment on the scene image in front of the user, performing super-resolution fusion reconstruction, and constructing a super-resolution scene image; performing object visual identification on the super-resolution scene image, and performing object bounding box segmentation processing to obtain a plurality of accurate scene object bounding boxes; and carrying out optical flow motion change analysis on the plurality of accurate scene object bounding boxes, and carrying out space-time scene structure evolution analysis so as to obtain dynamic scene evolution characteristic data. According to the method, the data processing efficiency of the environment perception data is improved, and the environment perception data is visualized more visually.

The invention discloses a terrain-combined weather radar base data dynamic resolution three-dimensional structure reconstruction method, which realizes three-dimensional structure reconstruction through combination of a digital elevation model (DEM) and weather radar base data. The method comprises the steps of cutting DEM topographic data and dumping the DEM topographic data into grid data in a binary form, generating a bitmap as a mapping source, establishing an elevation angle index array based on weather radar base data, establishing an azimuth angle index array based on a conical scanning surface so as to quickly map an angle, determining a rectangular reconstruction area and establishing a three-dimensional grid, traversing the three-dimensional grid to obtain an elevation angle, an azimuth angle and a distance combination array of grid points so as to obtain grid point data, smoothing the grid point data, re-sampling terrain grid data, constructing a terrain grid surface, obtaining a mapping image of the terrain grid surface, and rendering a three-dimensional body; the method has the advantages that the method does not need to depend on online service, and dynamic high-precision three-dimensional structure reconstruction of radar echoes in a specified area can be realized while the interaction performance is ensured.

The invention discloses a machine vision-based wiredrawing steel plate surface inclusion defect detection method and system. The method comprises the following steps: S1, acquiring a multi-light-source image sequence on the surface of a wiredrawing steel plate and carrying out brightness standardization; s2, decomposing the image sequence after brightness standardization through a Gaussian-Laplacian pyramid, and performing Laplacian pyramid fusion and image reconstruction to obtain a fused image; s3, carrying out directivity analysis on the wiredrawing texture in the fused image, extracting main direction features of the wiredrawing texture, and obtaining a wiredrawing direction consistency feature map; s4, performing feature extraction and feature fusion on the fused image and the wiredrawing direction consistency feature map based on a double-flow feature extraction network, and performing global average pooling operation; and S5, constructing and training a defect classification network, and predicting the defect category of the surface of the wiredrawing steel plate. The problems existing in the existing wiredrawing steel plate surface defect detection technology can be overcome, and the defect of inclusion on the surface of the wiredrawing steel plate is accurately detected.

The present invention discloses a method and system for replicating digital human images based on NeRF, and relates to the technical field of digital human image replicating. The method comprises: collecting digital human images to obtain training data sets; expanding and generating training samples; reconstructing a model using the expanded digital human image data training data; extracting features from the results of digital human image replicating output by the autoencoder, and inputting the feature vector after feature extraction into the classifier model to perform digital human image data screening. The present invention improves the details and authenticity of the generated samples, makes the expanded data closer to the real collected images in terms of fidelity, enhances the adaptability and stability when screening digital human image data, makes the screening model more adaptable when facing diversified replica data, improves the classification effect, and finally ensures the high quality and consistency of the replica data.

The invention provides a curtain wall reverse modeling method and system based on point cloud data, and is applied to the field of modeling data processing. The method can timely discover the defects of the model and improve the overall quality by accurately identifying irregular regions and performing error analysis, enhances the smoothness and detail performance of the model by applying a curved surface fitting algorithm and a hierarchical structure reconstruction technology, enables the model to be closer to a real object, and improves the accuracy of the model through structural feature-oriented local refinement. Geometric complexity is improved in a targeted manner, and uniformly distributed new vertexes are inserted to reduce zigzag and unsmooth areas, thereby improving overall fluency and visual effect.

The invention belongs to the technical field of artificial intelligence and coal mine production, and particularly relates to a coal gangue detection method based on multi-angle perception and mixed scale Transform feature aggregation, which comprises the following steps: firstly, acquiring a real-time video image of a conveying belt through a multi-angle camera device, and fully covering each area on the conveying belt; secondly, performing feature extraction and aggregation on the video image by adopting a mixed scale Transform model, fully fusing feature information under different scales, and enhancing the recognition capability of the model on the coal gangue; and finally, the aggregation features are analyzed by using a coal gangue detection algorithm, and high-precision coal gangue detection is realized. Real-time video images of the conveying belt are obtained through the multi-angle camera device, all areas of the conveying belt are fully covered, and therefore any blind area and any dead angle are eliminated. Advanced image preprocessing technologies such as de-noising and image enhancement are combined, so that the definition and detail performance of the image are remarkably improved.

The invention discloses an image acquisition method and system for a dual-wavelength circular polarization camera, and the method comprises the steps: setting two independent nanosecond pulse laser sources, and respectively transmitting lasers with different wavelengths; respectively carrying out left-hand circular polarization and right-hand circular polarization treatment on the two lasers; integrating the two nanosecond pulse laser beams subjected to circular polarization treatment into one beam, wherein the two beams are completely overlapped and are coaxial; the integrated nanosecond pulse laser beams are irradiated to a target object, and reflected light is collected through a double-spectrum camera to form two images with different wavelengths; and distributing weights for the image according to the importance of the polarization information, and generating a comprehensive image by adopting an image fusion technology. According to the image acquisition method and system of the dual-wavelength circular polarization camera, two independent nanosecond pulse laser sources are arranged to emit short-wavelength laser and long-wavelength laser respectively, the two beams of laser are subjected to left-handed and right-handed circular polarization processing and then integrated into one beam, and efficient imaging in a complex environment is achieved.

The embodiment of the invention discloses a low-illumination image enhancement method and device based on an improved normalized stream. The method comprises the following steps: acquiring a low-illumination image to be enhanced; performing normalization processing on the to-be-enhanced low-illumination image to obtain a normalized spectrum; performing edge extraction processing on the normalized spectrum to obtain an edge spectrum; performing feature fusion on the to-be-enhanced low-illumination image, the normalized spectrum and the edge spectrum to obtain a feature fusion image; performing feature extraction on the feature fusion image by adopting a feature extraction module to obtain a feature extraction image; performing feature mapping on the feature extraction image by adopting a feature mapping module to obtain a feature mapping image; and inputting the feature mapping image into the normalized stream reversible network, and outputting a normal illumination image. By enhancing the low-illumination image through the method, the enhancement effect can be improved.

The invention discloses an animation motion capture device for a virtual reality technology, and particularly relates to the technical field of virtual reality, which comprises a data acquisition module for capturing and collecting user motion and environment information, and a data processing module for analyzing, converting and optimizing the acquired data. Through cooperative work of the data acquisition module, the data processing module, the rendering feedback module and the like, accurate capture and real-time feedback of user actions are achieved, it is ensured that physical attributes of a virtual environment and a real environment are highly consistent, the data acquisition module can obtain multi-dimensional information such as somatosensory, expressions and environments of a user at high precision, and the user experience is improved. And the data processing module accurately identifies and optimizes user actions through a deep learning algorithm and a skeleton mapping technology, and the rendering and feedback module generates vivid visual, tactile and sound effect feedback, so that high immersive experience is brought, and fluency and consistency of actions and feedback are guaranteed.

The invention relates to the technical field of image processing, and relates to an inspection fuzzy image recognition method and device based on a region target detection algorithm. The method comprises the following steps: acquiring an original image shot by a to-be-identified device; carrying out image preprocessing, and inputting the image into the trained region target detection network for processing; and carrying out deblurring processing by adopting a multi-Wiener filtering algorithm to obtain a reconstructed clear image, and finally carrying out image recognition. According to the invention, a region target detection network based on YOLO network modification is adopted, so that the to-be-deblurred target region in the equipment image can be accurately identified while the model is lightened; according to the method, a multi-Wiener filtering algorithm is adopted, so that the to-be-deblurred target area in the equipment image can be effectively deblurred; according to the method, the clear image can still be generated under the condition that the inspection robot moves, the definition and detail performance of the image are greatly improved, the recognition accuracy is further improved, and misjudgment caused by blurring is reduced.

The invention provides a coal mine underground video dust filtration monitoring method based on model fusion, and belongs to the technical field of coal mine underground video dust filtration monitoring. The technical problem to be solved is to provide the coal mine underground video dust filtration monitoring method based on model fusion. According to the technical scheme, video data of an underground operation area environment are collected in real time through a camera installed in an underground coal mine, the video data are uploaded to a server arranged in a roadway in real time, the server firstly intercepts image frames from the underground coal mine video, timestamp marking is conducted on the image frames, and the image frames are sent to the roadway; carrying out histogram equalization preprocessing on each image frame, and improving the brightness and contrast of the collected video image to realize image enhancement; the server further adopts a dark channel prior method to carry out data recovery on the video image, a fuzzy effect caused by dust on the image is removed, and the real definition of the image is recovered; the method is applied to coal mine underground video dust filtration monitoring.

The invention discloses a digital human image duplicating method and system based on NeRF, and relates to the technical field of digital human image duplicating, and the method comprises the steps: collecting a digital human image, and obtaining a training data set; expanding and generating a training sample; training a data reconstruction model by using the expanded digital human image data; and performing feature extraction on a digital human image duplicating result output by the auto-encoder, and inputting a feature vector after feature extraction into the classifier model to perform digital human image data screening. According to the method, the details and the trueness of the generated sample are improved, the expanded data are closer to the actually acquired image in the aspect of trueness, the adaptability and the stability when the digital human image data are screened are enhanced, the screening model is more adaptive when facing diversified duplicated data, the classification effect is improved, and the classification efficiency is improved. And finally, high quality and consistency of the duplicated data are ensured.

According to the pulsar candidate body generation method and system based on the improved DCGAN, pulsar candidate body images are generated through the improved DCGAN, channel information of different scales is extracted, and then channel attention weights are generated through convolution processing of different scales; multi-scale features of output features are extracted, and then a space attention map is obtained based on different pooling operation processing; local details and global features of the image can be enhanced, so that the overall quality and detail performance of the generated image are remarkably improved.

The invention belongs to the technical field of image data processing, and particularly relates to an intelligent tobacco leaf curing method and system based on image recognition, and the method comprises the steps: obtaining a to-be-cured tobacco leaf feature map and a cured tobacco leaf feature map, and carrying out the imaging evaluation, thereby obtaining an imaging quality evaluation value; analyzing the imaging quality evaluation value, carrying out imaging adjustment, obtaining an imaging characteristic standard value, obtaining a to-be-flue-cured tobacco leaf quality parameter set based on the to-be-flue-cured tobacco leaf characteristic graph, calculating the to-be-flue-cured tobacco leaf quality evaluation value, carrying out to-be-flue-cured tobacco leaf pre-screening based on the to-be-flue-cured tobacco leaf quality evaluation value, and matching a tobacco leaf first-flue-curing mode, image segmentation is carried out on the characteristic graph of the flue-cured tobacco leaves to form a plurality of research and judgment units, quality monitoring values of the flue-cured tobacco leaves are obtained, a tobacco leaf redrying judgment result is obtained by comprehensively analyzing a tobacco leaf first-drying mode and the quality monitoring values of the flue-cured tobacco leaves, and the tobacco leaf redrying judgment result is matched with a tobacco leaf redrying mode. According to the method, accurate image recognition and classification in the tobacco leaf curing process are realized, and the applicability and scientificity of curing mode matching are improved.

The invention relates to an infrared image generation method based on multi-scale perception. Comprising the steps that image data are acquired and preprocessed, and the image data comprise a visible light image and an infrared image; constructing an infrared image generator and an infrared image discriminator; constructing a visible light image generator and a visible light image discriminator; a deep learning method is utilized to train an infrared image generator, an infrared image discriminator, a light image generator and a visible light image discriminator based on acquired image data. And determining the trained infrared image generator as an infrared image generation model. The method can generate a high-quality infrared image from a visible light image, and can be widely applied to the fields of night vision monitoring, remote sensing detection and the like.

The embodiment of the application discloses a low-illumination image enhancement method and device based on an improved normalized flow, which comprises the following steps: acquiring a low-illumination image to be enhanced; performing normalization processing on the low-illumination image to be enhanced to obtain a normalized spectrum; performing edge extraction processing on the normalized spectrum to obtain an edge spectrum; performing feature fusion on the low-illumination image to be enhanced, the normalized spectrum and the edge spectrum to obtain a feature fusion image; performing feature extraction on the feature fusion image by using a feature extraction module to obtain a feature extraction image; performing feature mapping on the feature extraction image by using a feature mapping module to obtain a feature mapping image; and inputting the feature mapping image into a normalized flow reversible network to output a normal-illumination image. The low-illumination image is enhanced by the method, and the enhancement effect can be improved.

The present invention discloses a method and system for generating images from text based on spectral normalization and contrastive learning. The method includes performing contrastive learning iterative training on a generator including spectral normalization according to two different key semantic features and word-level text features respectively corresponding to two different text descriptions, using the generator to generate multiple images of different resolutions respectively corresponding to the two text descriptions in each iterative training; a discriminator evaluates the similarity between the multiple images of different resolutions generated in each iterative training and the text descriptions; iteratively optimizes the generator according to each similarity result, until the generator that has been trained is obtained when the similarity result meets the conditions, and uses the trained generator to generate a target image from the text description. The present invention uses spectral normalization and contrastive learning in the generator network to assist the model in more in-depth learning of the complex associations of the same image under different text descriptions, so as to significantly improve the quality of the generated image.

The invention relates to the technical field of animation rendering, in particular to an animation scene fixation point rendering system based on eye movement tracking, which comprises a semantic attention modulation module, a rendering area dynamic construction module and a layered sample distribution rendering module, the semantic attention modulation module generates an attention diffusion diagram of semantic modulation based on the time sequence semantic importance maps of the current animation frame and the preorder frame and the current instantaneous fixation point data; a rendering area dynamic construction module adaptively constructs a core rendering area and a gradual transition rendering area according to the attention diffusion diagram; and the layered sample distribution rendering module implements ray tracing sampling of different sample numbers on the core rendering area, the gradual transition rendering area and the background area on the GPU. Compared with the prior art, the image quality of the gazing area is improved under the limited computing power, and the overall visual transition is kept smooth.

The invention provides a near-eye light field scanning display method and device, electronic equipment and a storage medium, and the method comprises the steps: displaying a plurality of to-be-processed near-eye light field display contents at different moments based on a time division multiplexing technology; performing offset display processing on the plurality of near-eye light field display contents to be processed to obtain a plurality of offset near-eye light field display contents; the multiple pieces of shifted near-eye light field display content are stacked, target near-eye light field display content is obtained, and the first spatial resolution of the target near-eye light field display content is higher than the second spatial resolution of the near-eye light field display content to be processed. And the spatial resolution of the near-eye light field display content is improved.

The present invention relates to the field of optical detection, and in particular to a rescanning confocal measurement system and a measurement method based on dynamic focus weighted fusion, wherein the excitation light emitted by a laser light source is sequentially reflected by a dichroic mirror, swung by a first scanning oscillating mirror, focused by a first scanning lens, aberration corrected by an aberration correction lens, and focused by an objective lens before being incident on a sample to be tested, and the emission light excited by the sample to be tested returns to the dichroic mirror, is transmitted by the dichroic mirror, focused by a second scanning lens, passes through a pinhole, and then collimated by a third scanning lens, swung by a second scanning oscillating mirror, and split by three beam splitters, and is incident on two pre-focus detectors and two post-focus detectors, respectively, to obtain a total of four defocused images. The present invention introduces a rescanning unit and an image fusion method, so that images can be directly projected onto cameras with four different focal planes, thereby enhancing the details and clarity of the image, reducing blur and out-of-focus areas, thereby reducing the impact of noise, and improving the overall quality of the image.

The present invention discloses a high-quality audio generation method for a DAS system, which pre-processes and extracts features of low-quality and high-quality audio data, and obtains data sets D1 and D2 respectively, which are a collection of pitch feature-Mel spectrum feature data pairs of multiple short time frames; constructs a high-quality audio generation model including a generative network and a discriminant network with multi-channel input; uses D2 to train the generative network to learn the prior distribution of high-quality audio; the discriminant network judges the generated signal of the generative network, and if the loss function is less than a set threshold, the training is terminated, otherwise the generative network and the discriminant network are optimized and the training is repeated; then the high-quality audio generation model is trained with D1 and D2; after pre-processing and feature extraction of the low-quality audio to be processed, the trained high-quality audio generation model is input to obtain high-quality audio. The present invention improves the spatial resolution and consistency of audio signals, and can generate high-quality audio more accurately.

The present invention discloses a method and system for detecting defects of inclusions on the surface of a brushed steel plate based on machine vision, including: S1: obtaining a multi-light source image sequence of the surface of a brushed steel plate and performing brightness standardization; S2: decomposing the image sequence after brightness standardization through a Gaussian-Laplacian pyramid and performing Laplace pyramid fusion and image reconstruction to obtain a fused image; S3: performing directional analysis on the brushed texture in the fused image, extracting the main direction features of the brushed texture, and obtaining a brushed direction consistency feature map; S4: performing feature extraction and feature fusion on the fused image and the brushed direction consistency feature map based on a dual-stream feature extraction network, and performing a global average pooling operation; S5: constructing a defect classification network and performing training to predict the defect category on the surface of a brushed steel plate. The present invention can overcome the problems existing in the existing brushed steel plate surface defect detection technology and accurately detect defects of inclusions on the surface of a brushed steel plate.

The present invention relates to the technical field of model optimization, and in particular to a method and system for optimizing a large model of a Wensheng graph based on multi-objective optimization. The method comprises: acquiring text data; constructing a large model of a Wensheng graph based on a diffusion model; constructing a multi-objective optimization problem based on optimization variables, optimizing parameters of the large model of the Wensheng graph based on the multi-objective optimization problem and using the NSGA‑II algorithm; based on parameter optimization, using the large model of the Wensheng graph to perform reverse diffusion to generate a final user demand image. The Wensheng graph optimization fine-tuning method proposed in the present invention can optimize multiple conflicting objectives at the same time. For the fine-tuning of the Wensheng graph model, the clarity, color fidelity and detail expression of the image can be improved at the same time. At the same time, it is ensured that the generated image is highly consistent with the input text in semantics.

The present invention discloses a method for detecting multiple types of defects on the inner wall of a corrugated pipe based on deep learning, comprising the following steps: S1, collecting an initial image of the inner wall of the corrugated pipe and images of the inner wall of the corrugated pipe under different lighting environments to generate an original image; S2, preprocessing the original image to generate a preprocessed image; S3, extracting the overall surface geometric features of the inner wall of the corrugated pipe from the preprocessed image through a pre-trained ResNet101 network; S4, inputting the preprocessed image and the overall surface geometric features into an improved StyleGAN3 network to generate an enhanced image; S5, inputting the enhanced image into a multi-scale adaptive deep network to output a defect detection result; S6, post-processing the defect detection result to output a final defect detection result. The present invention adopts deep learning and geometric feature embedding technology to realize accurate detection of defects on the inner wall of the corrugated pipe, and has the advantages of high precision, strong robustness and comprehensive adaptability.

The invention relates to the technical field of image processing, and relates to an instrument image super-resolution reconstruction method and device based on NLSA and SWL. The method comprises the following steps: acquiring a low-resolution instrument image XLR; performing image enhancement preprocessing on the XLR to obtain an enhanced image XEm; and processing the XEm through a trained super-resolution reconstruction model to obtain a reconstructed high-resolution instrument image XHR. According to the invention, a self-designed super-resolution reconstruction model is adopted, on one hand, multiple attention mechanisms are fused to realize multi-scale and all-around feature extraction, and on the other hand, a loss function including static wavelet loss and L1 loss of an RGB domain is adopted to ensure a model training effect; therefore, real-time reconstruction is achieved under the condition that the model reconstruction effect is guaranteed, and the problem that an existing super-resolution reconstruction method based on the CNN or Transform model is not suitable for processing cabinet instrument images is solved.

The invention relates to a system for detecting a focal plane of an infinite-distance microscopic system. The system comprises an objective lens, a spectroscope, a tube lens, an image acquisition module and an image processing module. The objective lens collects light reflected by an object and converts the light into parallel light, and the spectroscope separates the parallel light into two parts which are respectively transmitted to the tube lens and the image acquisition module. The first tube lens receives part of light and focuses the light on an image focal plane, the second tube lens in the image acquisition module receives the other part of light and focuses the light on the digital camera photosensitive chip, and the digital camera photosensitive chip is obliquely mounted to ensure light focusing of a specific field of view. And the image processing module processes the digital image through partitions and analyzes the definition curve graph, and calculates an object plane defocusing amount to determine an optimal focal plane position.