175 results about "Image boundary" patented technology

The invention relates to the technical field of topographic change analysis, in particular to an unmanned aerial vehicle-based topographic change detection system, which comprises a slope direction sensing track control module, a texture structure extraction module, a crack evolution track construction module, a direction trend comparison module and a patrol recheck positioning module. According to the method, a continuous elevation point column of an unmanned aerial vehicle scanning area is extracted, laser reflection point coordinates are fused, a space relation of transition point distribution is constructed, dynamic adjustment of a ground-imitated flight path is achieved, and a texture structure area with continuous directivity is recognized in combination with a high-angle image boundary communication relation; texture boundary evolution is compared at different time nodes to form a crack path, the stability of the path and the slope direction is judged through an included angle sequence, recognition and sorting of areas with the consistent direction are completed, a space comparison result is registered in a three-dimensional coordinate system, terrain change areas are accurately marked, and rapid positioning and continuous tracking of high-risk areas are achieved.

The invention discloses an image edge cutting method and device based on boundary detection, equipment and a storage medium, and relates to the technical field of image boundary detection, and the method comprises the steps: carrying out the preprocessing of an original image, and obtaining a target image; determining a scene type of the target image; based on an edge detection strategy corresponding to the scene type, performing edge detection on the target image to obtain an edge image; performing straight line detection on the edge image to obtain a plurality of candidate straight line segments; and based on a boundary positioning rule, determining a target boundary line segment from the plurality of candidate straight line segments, so as to carry out edge cutting on the original image according to the target boundary line segment. According to the invention, a more reliable image edge cutting effect can be generated.

The invention provides an image region-of-interest extraction method and system based on a Mama architecture, and relates to the technical field of image processing, and the method comprises the steps: obtaining a to-be-extracted image; performing multi-scale feature extraction on the to-be-extracted image through a local enhancement module; multi-scale semantic enhancement features are generated through a cross-scale self-attention module, and dimension reduction processing is performed through a feature conversion module; local detail features are generated through an adaptive detail enhancement module; performing global context enhancement processing on the local detail features through a pyramid pooling module to generate context enhancement features; performing up-sampling processing on the context enhancement features; the context enhancement features after up-sampling processing are fused through a self-adaptive global-local fusion gating module; and carrying out image extraction based on the decoded fusion features. The image segmentation precision is improved, the model is light in weight, reasoning is fast, and the problems of image boundary blurring and scale variability are effectively solved.

The invention relates to the technical field of anomaly detection, in particular to a circuit board processing anomaly detection system which comprises an image anchoring module, a gray scale comparison module, a candidate area screening module, a gradient closing verification module and a behavior track evolution module. According to the method, by setting the gray scale anchor points in the image frames and collecting the standard gray scale values, deviation accumulation caused by the overall illumination change of the image can be effectively avoided, the response delay area is recognized through inter-frame gray scale comparison, and early intervention on the etching abnormal process is achieved; a gray scale fluctuation low-amplitude segment in an image boundary region is extracted to recognize a texture missing region, the recognition accuracy of structural blur is enhanced, fine extraction of an abnormal structural form is realized in combination with gray scale gradient chain construction and a closed error judgment mechanism, region response change trajectory tracking in a frame sequence is superposed, and the recognition accuracy of structural blur is improved. Therefore, the judgment is not limited to single-frame anomaly detection and has a time sequence evolution recognition capability, an abnormal symptom region is stably captured, and a reliable detection result is generated.

Systems and methods for image processing (e.g., image extension or image uncropping) using neural networks are described. One or more aspects include obtaining an image (e.g., a source image, a user provided image, etc.) having an initial aspect ratio, and identifying a target aspect ratio (e.g., via user input) that is different from the initial aspect ratio. The image may be positioned in an image frame having the target aspect ratio, where the image frame includes an image region containing the image and one or more extended regions outside the boundaries of the image. An extended image may be generated (e.g., using a generative neural network), where the extended image includes the image in the image region as well as generated image portions in the extended regions and the one or more generated image portions comprise an extension of a scene element depicted in the image.

The invention relates to the technical field of wearable monitoring, in particular to farm intelligent wearable inspection equipment, which comprises a track thermal difference recognition module, an abnormal target labeling module, a high fever behavior screening module, an image information extraction module, a structure label generation module and an individual inventory recognition module. According to the invention, by extracting the infrared temperature measurement data and calculating the thermal fluctuation difference value, the positioning of a temperature abnormal section is realized, the identification capability of tiny thermal environment variation is enhanced, the high-heat target area is positioned, the image filing is completed, and the target screening accuracy and the image utilization efficiency are improved. Static high-heat individuals are identified based on the calorific value state and the spatial position stability, the discovery probability of potential abnormal objects is improved, through image boundary contour extraction and structural feature comparison, the individual identification efficiency is improved, and image information and dynamic path data are synchronously analyzed, so that the monitoring result is more continuous and complete, and the detection accuracy is improved. And the intelligent response capability and the information processing depth are improved.

The invention provides an urban digital twinning modeling method and system based on multi-source heterogeneous data fusion, and relates to the technical field of digital twinning, and the method comprises the steps: obtaining a registration image and a registration point cloud of a target region, respectively extracting an image boundary candidate and a point cloud boundary candidate, generating a boundary uncertain band for each boundary candidate, and carrying out the boundary uncertain band extraction; determining an uncertain band width and an offset direction identifier; boundary matching pairs are formed according to the spatial intersection relation, and intersection sections with opposite bias direction identifications are determined as a confrontation boundary section set; for each confrontation boundary section, determining an image confidence weight and a point cloud confidence weight, and generating a boundary judgment result according to the image confidence weight and the point cloud confidence weight; and performing topological consistency verification on a boundary judgment result, and determining a unique fusion boundary. According to the method, under the condition that local physical characteristic difference and systematic geometric deviation exist in multi-source data, the geometric edge fusion precision and topology reliability can be improved, and the stability and availability of the urban digital twin model are enhanced.

The invention discloses a double-flow fault diagnosis method based on state space modeling and dynamic edge context awareness, the method constructs a time domain state flow and a time frequency edge flow, and the specific implementation process is as follows: the time domain state flow is subjected to state space modeling under a Transform framework to form a MambaT encoder, long and short term dependence is efficiently captured with linear complexity, and the time domain state flow is subjected to dynamic edge context awareness; time domain sensitive features are accurately represented; the time-frequency edge stream converts vibration data into a two-dimensional time-frequency image through wavelet transformation, a differentiable edge sensitive mask is designed to adaptively separate time-frequency image impact characteristics and background noise, boundary content perception adaptive filling is designed to maintain image boundary topology integrity, and the two parts cooperate to form dynamic edge context perception convolution, so that the image boundary topology integrity is improved. Time-frequency edge features are effectively extracted; and a channel enhancement and channel gating module is designed to establish a dynamic channel interaction gating mechanism, so that time domain-time frequency domain double-current cascade characterization is realized, and the fault discrimination capability is remarkably improved. Experiments prove that the method has higher fault diagnosis performance.

The invention discloses a sleeve grouting quality detection method and system based on multi-modal data fusion, and belongs to the technical field of grouting quality detection.The method comprises the steps that a visual image sequence, an acoustic emission signal sequence and local temperature dot matrix data are synchronously collected in the grouting process, and a multi-modal input set is formed; image boundary change, acoustic jump features and a thermal diffusion track are extracted respectively, a fusion feature vector is constructed and input to a space-time heterogeneous coupling network, and a local grouting dynamic anomaly map is generated; then, residual confidence fusion analysis is carried out, a defect scoring graph is output, and intelligent judgment of defect positions and types is achieved in combination with grouting flow direction topological information; the method can realize accurate identification, classification and visual positioning of grouting defects, and is suitable for scenes such as quality monitoring of an assembly type structure.

The invention discloses an ultrasonic image boundary perception segmentation method, system and device, and relates to the field of image processing, and the method comprises the steps: obtaining an ultrasonic image, and sequentially processing the ultrasonic image through a block embedding module and an image coding module to obtain spatial domain features; performing frequency domain feature extraction processing on the spatial domain feature by using a preset multi-scale frequency extraction strategy to obtain a fine-grained high-frequency feature, a coarse-grained high-frequency feature and a low-frequency feature, and obtaining a frequency fusion feature in combination with a preset frequency alignment strategy; a preset frequency guide boundary refining strategy is combined to determine refining features; and according to the refined features and a preset boundary guiding decoding strategy, determining a segmentation mask representing a result of segmenting the target region under the ultrasonic image from the background region. A low-frequency structure and multi-scale high-frequency boundary details are explicitly separated through frequency domain decomposition, an ultrasonic image boundary sensing segmentation scheme based on frequency guidance is provided, the boundary sensing ability is high, and the cross-domain generalization ability is high.

According to the image calibration and splicing method based on sub-pixel reconstruction and elliptic coefficient weighting, a super-resolution reconstruction module is introduced before image array splicing, the splicing precision is improved through sub-pixel-level image registration, and the splicing precision better than that of a single pixel unit can be achieved. Aiming at the problems of edge dislocation and stiff transition existing in multi-camera array system view field splicing, the invention provides a weighted splicing algorithm based on an elliptic coefficient, and the algorithm constructs an elliptic gradient weight model, realizes smooth transition of an image boundary region, and remarkably improves the overall visual effect of a panoramic image.

The invention relates to the technical field of multi-modal fusion, in particular to an automatic driving complex scene sensing method based on multi-modal fusion, which comprises the following steps of: acquiring an image bounding box and a point cloud cluster center position, identifying and pairing space overlapping targets, analyzing direction change to extract a track difference frame segment, and extracting a track difference frame segment; the edge trend and point cluster arrangement are linked to divide modal sources, contact areas and modal attribution are concluded according to side lines, and a perceptual path sorting framework is constructed according to an arrangement fusion sequence. According to the method, a cross-modal target association relationship is constructed through spatial overlapping pairing of an intra-frame image and a point cloud modal target, track difference fragments are extracted through a continuous frame direction change sequence, behavior change positions are distinguished, and modal sources are divided through conjoint analysis of an edge trend and a point cluster arrangement trend; and establishing an association matching relationship between a modal and a scene component through extraction of a contact area, arranging a fusion sequence through contact expression of the modal in a difference frame segment, and constructing a perceptual path sorting framework with distinguishing capability.

The invention discloses a steel structure drilling deviation real-time correction method based on visual identification, and relates to the technical field of drilling deviation analysis, and the method comprises the steps: collecting a steel structure image before drilling, and calculating an image center coordinate value and image boundary included angle information of a current drilling target position based on a hole position contour and an edge line segment in the image; constructing a deviation reference set containing the historical drilling task number, the image center coordinate value and the image boundary included angle information, and based on the current image center coordinate value and the image boundary included angle information, retrieving a plurality of spatially adjacent reference records from the deviation reference set to generate a first local deviation estimation value; an image sequence before feeding of the drill bit in the drilling process is collected, and a second local deviation estimated value is output based on a local image block of a contact area between the edge of a drill bit cone and the surface of the component in the image sequence; real-time deviation correction and path continuity control in the steel structure drilling process are achieved.

The invention discloses a point cloud and image fused robot structure modeling method and system. Firstly, the invention provides an image boundary guided point cloud clustering method, image edge and region consistency features are fused, edge consistency loss and clustering contour score are taken as a joint optimization target, and the problems that pure point cloud clustering is fuzzy in component boundary and easy to merge in texture similar structures are solved; and the separation precision of key connecting parts is obviously enhanced. Secondly, the invention innovatively provides a motion consistency analysis mechanism driven by a time sequence and an image, carries out linkage optimization on a spatial clustering result under the guidance of the image and a multi-frame motion track, constructs a motion correlation graph, and deduces the structural topology of the robot based on a minimum spanning tree. And high-robustness component connection relation restoration is realized under the condition of no label supervision. The strategy is excellent in performance on simulation and physical data sets, and the topology accuracy is improved by 6.2%.

The invention discloses a dual-energy ray bone mineral density evaluation method based on a flat panel detector, and relates to the technical field of ray evaluation. Through a direct conversion mode of the flat panel detector, high-precision acquisition of low-energy and high-energy ray images is realized, and noise and resolution loss caused by traditional indirect imaging are remarkably reduced; primary correction and secondary correction are completed in the edge calculation nodes, so that the problems of discontinuous gray scale and distorted energy ratio of the boundary area of the skeleton and the soft tissue are effectively solved, and the authenticity of the image boundary is improved; and dynamic smoothing processing of the energy ratio is further combined to obtain a stable ratio image, so that the stability and repeatability of bone mineral density measurement are remarkably enhanced, and more accurate bone mineral density evaluation is realized.

This disclosure provides a video processing method and apparatus. An example method includes: determining whether a coded block includes samples outside an image boundary; and, in response to the coded block being determined to include samples outside an image boundary, performing quadtree segmentation of the coded block regardless of the value of a first parameter, wherein the first parameter indicates whether the quadtree is permitted for segmenting the coded block.

A high-robustness mixed data enhancement synthesis method, comprising the steps of: 1. preprocessing the input original image to generate an accurate binary mask; 2. performing transformation enhancement on the synthesized image, including but not limited to rotation, shearing, erosion and dilation operations; 3. through a probability-weighted random allocation mechanism, randomly generate high-quality irregular detection targets in stress-concentrated positions in terms of the number of detection objects and the physical layer, simulate real fault scenarios; 4. using a matching algorithm to paste the samples of the detection targets onto the background image, and introducing a boundary overflow processing mechanism to ensure that the detection targets do not exceed the ROI or image boundary, and to ensure that the position of the detection target corresponds to the image, and to ensure that the label does not change. This method is applied to application scenarios with scarce data sets in industrial scenarios, significantly improving the diversity and complexity of the data set, thereby greatly improving the generalization ability and detection robustness of the subsequent target detection model in complex environments.

The application relates to an object boundary recognition method, device and equipment, which comprises the following steps: recognizing a target object in an input image to obtain a target object image; filling the target object image into a square, performing binary image segmentation on the target object image to obtain a binary image; performing full-image boundary processing on the input image to determine a full-image boundary; connecting intersection points obtained by intersecting the full-image boundary and the boundary of the binary image to obtain the boundary of the target object. The application can be more accurate in picture boundary prediction, and manual secondary modification is easy to operate because the prediction result can be ensured to be more accurate, and the inaccurate part is convenient for manual boundary supplement.

The application discloses a kind of high-efficiency boundary detection methods based on angle entropy weighted spring network, it is related to image boundary detection technical field, the method includes: for each data point in data set, search its multiple near neighbors;Calculate the elasticity of each near neighbor to current data point, elasticity is based on Hooke's law simulates spring connection;The resultant force of all elasticities is calculated as basic boundary score;Calculate the angle entropy of current data point, angle entropy is based on the angle information of near neighbor distribution Quantification direction distribution;With angle entropy weighted basic boundary score to obtain final boundary score;Select boundary point by pre-set threshold, complete boundary detection.The application quantifies local stress imbalance by spring network, and utilizes angle entropy to amplify boundary feature, while ensuring linear time complexity, significantly improves the discrimination of boundary point detection.

This application provides a method for adding watermarks to an image. Belonging to the field of information security, this method involves dividing an image into blocks to obtain a target number of non-overlapping blocks. The texture complexity is determined based on the number of pixels at the block boundaries and the size of the blocks. The watermark embedding amount is determined based on the texture complexity of each block, the average watermark embedding capacity, and the initial texture complexity. The average watermark embedding capacity is determined based on the total watermark information capacity and the target number. Singular value decomposition is performed on each block to obtain the maximum singular value. The singular value after watermark embedding is determined based on the maximum singular value of each block and the watermark information corresponding to the watermark embedding amount. Inverse singular value decomposition is performed on the singular values ​​after watermark embedding in each block, and each block after watermark embedding is reconstructed to obtain a watermarked image. This method can improve the amount of watermark information and the quality of the watermarked image.

The invention discloses a highway meteorological data acquisition method and system based on meteorological bulletin images, and belongs to the technical field of meteorological data processing and geographic information systems. The method comprises the following steps: acquiring a meteorological bulletin image from the China Meteorological Bureau, and extracting effective meteorological data through preprocessing operations such as image boundary determination, scale-pixel conversion and data denoising; converting image pixel coordinates into latitude and longitude coordinates, and superposing the latitude and longitude coordinates with a road vector map to generate a road weather chart; identifying meteorological values based on a color threshold method, and extracting a longitude and latitude sequence of a target road section in combination with a region growing algorithm; and calculating continuous meteorological data by adopting a point-by-point interpolation method, and finally displaying meteorological distribution along the highway through a visualization technology. The problems that an existing meteorological data acquisition mode is high in cost, complex in operation and low in reliability are solved, low-cost and high-precision automatic extraction of the highway meteorological data is achieved, and real-time data support is provided for traffic management and disaster early warning. The system can be widely applied to the fields of highway meteorological monitoring, geological disaster early warning and the like.

The invention relates to an obstacle truncation attribute marking method and device, equipment and a medium, and the method comprises the steps: collecting obstacle information through a radar sensor and an image sensor, and obtaining a detection frame of an obstacle in an image; for a target obstacle in which a detection frame of the obstacle in the image and an image boundary have a truncation relationship, converting the detection frame of the target obstacle to a preset view angle; acquiring a first geometric parameter corresponding to a part, outside an image sensor FOV, of a detection frame of the target obstacle and a second geometric parameter corresponding to the detection frame of the target obstacle at the preset view angle; and generating a truncation attribute of the target obstacle through a ratio between the first geometric parameter and the second geometric parameter, so as to mark the image by using the truncation attribute of the target obstacle. According to the technical scheme of the invention, automatic marking of the obstacle truncation attribute can be realized.

To provide an apparatus and a method for dividing an image into coding units.SOLUTION: An image is divided into hierarchically partitioned coding tree units (CTUs). Hierarchical partitioning includes multi-type partitioning, such as binary tree or quad tree partitioning. For CTUs that are entirely within the image and CTUs on the boundary, respective multi-type partition depths are selected. The present disclosure provides the flexibility of multi-type partitioning in the boundary portion of the image.SELECTED DRAWING: Figure 10

The application discloses a kind of based on pattern recognition's storage tank corrosion defect C scan analysis method, including the following steps: obtaining storage tank C scan image, decay characteristic time chart, amplitude chart and wall thickness estimation chart;With the preprocessed data set input image boundary passage of cross-modal corrosion boundary perception network;With the preprocessed data set input physical decay passage;Based on the cross-modal fusion module of cross-modal corrosion boundary perception network, cross-modal attention fusion is executed;With the cross-modal fusion feature set input boundary refinement module, generate corrosion region initial segmentation chart and corrosion boundary confidence chart;Based on corrosion boundary confidence chart, corrosion region initial segmentation chart is executed iterative boundary refinement processing;According to target corrosion region segmentation chart and preprocessed data set, establish spatial correspondence.This application cross-modal corrosion boundary perception network realizes the fine identification and quantitative analysis of storage tank corrosion region.

A block partitioner includes a quad splitter structured to partition a target block obtained by recursive partitioning in half in both a horizontal direction and a vertical direction to generate four blocks, and a binary / ternary splitter structured to partition the target block obtained by recursive partitioning into two or three in the horizontal direction or the vertical direction to generate two or three blocks, and the binary / ternary splitter disallows partitioning of the target block in the horizontal direction when partitioning of the target block in the horizontal direction causes the target block obtained by partitioning to be located beyond a right side of a picture boundary, and disallows partitioning of the target block in the vertical direction when partitioning of the target block in the vertical direction causes the target block obtained by partitioning to be located beyond a lower side of the picture boundary.

The invention provides an engineering specification document processing method and system, and relates to the field of data set production, and the method comprises the steps: generating a chapter index sequence based on the hierarchical number characteristics of an engineering specification document; the method comprises the following steps: converting an engineering specification document into a high-resolution image, identifying text content based on an OCR technology, extracting a numbering mode from the OCR identification content through a regular expression, and accurately matching the numbering mode with a chapter index sequence so as to obtain page numbers and coordinate information of identified chapters; the missing position is calculated according to the coordinate information of the adjacent chapters, and interpolation completion is carried out on the missing coordinate information; calculating the image boundary of each chapter according to the complemented coordinate set, cutting the picture of the engineering specification document according to the image boundary, setting a boundary protection mechanism in the cutting process to maintain the integrity of the text content of each chapter, and finally forming a structured image library which can be directly used for knowledge management and intelligent retrieval. And the engineering specification retrieval efficiency and accuracy are improved.

The invention provides a bacterial colony image segmentation method, system and device based on AI, and a medium, and belongs to the technical field of bacterial colony selection, and the method comprises the steps: building a bacterial colony image collection environment; the method comprises the following steps: placing a culture dish with a bacterial colony to be detected in a bacterial colony image acquisition environment, acquiring an RGB image of the culture dish, extracting an effective area of the culture dish, removing background interference, segmenting the effective area into sub-graphs with the same size, and setting a sub-graph overlapping area; reasoning each sub-graph by using a pre-trained bacterial colony detection model, and outputting a mask and a bounding box of a bacterial colony; combining reasoning results of the sub-graphs, and deleting target bacterial colonies of which the distance from a central point to a sub-graph boundary is smaller than a set pixel threshold value; and calculating the diameter, roundness and proximity of the bacterial colonies based on the masks of the bacterial colonies, and screening out the bacterial colonies meeting conditions based on a preset threshold value. The accuracy and integrity of bacterial colony selection are improved, and the selection efficiency is improved.

This application discloses a fisheye image correction method applied to an in-vehicle stitching system. The method includes acquiring a fisheye image F using an in-vehicle fisheye lens; determining the edges of black borders within the fisheye image F; performing curve fitting on the pixels at the edges of the black borders; mapping the fitted curve to the image boundary of the fisheye image F to obtain a correction mapping relationship; and acquiring a fisheye image T to be corrected using the in-vehicle fisheye lens, and performing mapping correction on the corrected fisheye image T using the correction mapping relationship. This application also provides a computer-readable storage medium.

The invention relates to the technical field of intelligent information processing, in particular to an enterprise financial report authenticity verification method fusing OCR and financial logic verification, and the method comprises the steps: carrying out the preprocessing and structure analysis of a financial report image file, generating a regionalized image set with metadata, carrying out the feature extraction of the regionalized image set, and carrying out the verification of the authenticity of the financial report. The method comprises the following steps: obtaining high-dimensional visual features, carrying out serialized decoding on the high-dimensional visual features, outputting a character probability sequence, decoding the character probability sequence to obtain a corresponding identification value, binding the identification value with bounding box coordinates, identification confidence and accounting subject identifiers of the identification value in an original financial report image, and constructing a structured financial data unit; and a multi-dimensional fusion verification mechanism is synchronously triggered, and finally, various verification results are comprehensively analyzed, and financial report authenticity risk rating is generated, so that deep fusion of financial report image identification results and financial logic rules is realized, and the automation degree and risk identification capability of enterprise financial report authenticity verification are effectively improved.