109 results about "Boundary contour" patented technology

The invention provides a time-space evolution data processing method and system applied to tunnel deformation analysis, and the method comprises the steps: obtaining a time sequence monitoring sequence of different monitoring dimensions in a tunnel construction process, constructing a time-space evolution correlation model of tunnel deformation based on the time sequence monitoring sequence, and analyzing a dynamic coupling relation between structure response data and geological environment action data, generating a space-time correlation feature matrix, generating a feature evolution sequence of tunnel deformation based on the space-time correlation feature matrix, and performing nonlinear mutation feature extraction on the feature evolution sequence to obtain a mutation feature index set; according to the sudden change characteristic index set and a preset accumulated deformation rate threshold value, performing dynamic division of tunnel deformation safety levels, generating a safety level division result, and based on the safety level division result and the spatial distribution data of the tunnel construction section, executing spatial positioning and boundary contour extraction of the potential risk damage area. And obtaining a risk damage area distribution map. According to the invention, the tunnel construction safety risk can be accurately evaluated and effectively controlled.

The invention relates to the technical field of image segmentation, in particular to an intensive care unit video image processing method based on image semantic segmentation, which comprises the following steps of: acquiring a video image through monitoring camera shooting, dividing a semantic region, calculating a pixel displacement direction of each part, identifying an abnormal track, erasing interference, recombining a boundary contour, and comparing shape change. And constructing a trend and re-drawing a structure, evaluating stability in combination with directions, speeds and amplitudes, screening and correcting inconsistent labels, and outputting an index result. According to the method, track features are constructed by introducing pixel direction coding, a deviation region is identified and interference is eliminated in combination with a direction change trend, an edge structure is reconstructed by using a boundary connection sequence and curve fitting, contour division is optimized through deformation direction consistency, and labels are corrected by synthesizing direction change and boundary speed. Dynamic tracking and accurate labeling of the patient state are achieved, the abnormal action recognition efficiency and the label updating accuracy are improved, and the time sequence integrity and the space expression ability of image data are enhanced.

The invention discloses a visualization method and system for an oversized in-situ filling space under a composite roof, and belongs to the technical field of information processing. The method comprises the following steps: performing multi-angle data acquisition on a filling space under a composite roof structure to obtain original point cloud data; obtaining complete three-dimensional point cloud data based on the original point cloud data; preprocessing the three-dimensional point cloud data to obtain standard point cloud data; performing boundary contour recognition based on the standard point cloud data to obtain space boundary contour data; obtaining a spatial three-dimensional geometric model based on the spatial boundary contour data; performing section slicing based on the space coordinate range of the space three-dimensional geometric model, and obtaining size distribution data based on the obtained contour size parameters of the section; obtaining a total volume and a segmented volume of the filling space based on the size distribution data; and obtaining a three-dimensional visual model of the filling space based on the total volume and the segmented volume. The method provides reliable technical support for mine filling design and resource evaluation.

The invention discloses a Thangka image target detection method based on improved YOLOv11. The method comprises the steps of S1, acquiring a traditional art data set and performing preprocessing; s2, a backbone network part is improved based on a YOLOv11 network model, specifically, multi-scale edge feature enhancement conduction module design is added, and the sensitivity of the model to a boundary contour and the positioning precision of a target in a complex structure are improved; s3, the neck network part is improved based on a YOLOv11 network model, specifically, a windmill convolution module is adopted to replace traditional down-sampling convolution, a direction perception and asymmetric convolution strategy is adopted, and a receptive field is expanded on the premise that significant calculation overhead is not increased; s4, a detection head part is improved based on a YOLOv11 network model, specifically, a lightweight scale adaptive detection head is adopted, a convolution sharing mechanism and SimAM non-parameter attention are fused, and the parameter quantity of the detection head is reduced; and S5, according to the improvements in S2, S3 and S4, integrating into an improved model after structure optimization.

Disclosed in the present invention is a three-dimensional circle detection method for a freeform surface, comprising: acquiring freeform surface point cloud data, and using a normal vector-guided rolling ball method to extract a freeform surface circular-hole point cloud boundary contour from the point cloud data; performing Euclidean clustering-based boundary segmentation on the point cloud boundary contour to obtain a clustering result of all boundary point clouds, the clustering result consisting of a plurality of circular-hole contours; performing a density-based weighted iterative algorithm on each circular-hole contour to obtain a freeform surface circular-hole normal vector; and projecting a freeform surface circular hole along the freeform surface circular-hole normal vector onto a plane, and then performing circular-hole iterative fitting based on solving an overdetermined equation on the projected circular-hole point cloud to obtain final positioning information. The present invention is applicable to circular-hole positioning on freeform surfaces of different curvatures, has strong robustness, and effectively improves the accuracy of circular-hole positioning on the freeform surfaces.

The invention relates to the technical field of workpiece alignment, in particular to a machined workpiece automatic alignment method and system based on image measurement, and the method comprises the following steps: collecting a workpiece image, analyzing a local curvature change rate, adjusting sampling density, reconstructing surface features, generating micro-topography data, calculating a stability score, and outputting an attitude correction parameter. And tracking a time sequence deviation derivation instruction response deviation track, fitting a deviation curve to invert delay mapping to generate an error correction table, and mapping a control signal to adjust a step length to generate an adaptive driving instruction. According to the method, the morphological stability score is calculated through the ratio of the boundary curvature continuity to the area compactness of the equal-optical-density region, accurate judgment on the boundary contour under different image conditions is achieved, the response offset function is constructed with the help of the least square method, the execution error correction table is mapped, the response precision and execution consistency of the alignment instruction are improved, and the accuracy of the alignment instruction is improved. And the alignment efficiency and the posture stability of the workpiece under the complex morphology condition are improved.

The invention relates to a method and system for identifying a projectable area of a projector. The method comprises the steps of establishing a coordinate system taking the center of a projector lens as a base point based on boundary contour parameters of the projector, generating corresponding contour line data by utilizing the boundary contour parameters, and projecting contour lines on a projection surface to identify a movable axis range and / or a variable focus range of the projector lens. The system comprises a projection optical module, a boundary detection unit, a contour generation unit, and an independent contour projection unit and a control unit which are arranged on a shell. According to the scheme, the contour line data is generated based on the boundary contour parameters and is projected on the projection surface, so that a user can intuitively perceive the adjustable limit range, blind adjustment is avoided, the risks of image distortion, dark corners and content missing are reduced, the adjustment efficiency and accuracy are improved, practicability and flexibility are both achieved, and the user experience is improved. The usability of the projector and the user interaction experience can be effectively improved, and the defect of lack of visual boundary prompt in the prior art is overcome.

The invention discloses a full-automatic measurement and strength testing method and system for a concrete core sample, and the method comprises the steps: obtaining an end face image and a cylindrical surface image of the concrete core sample, and extracting the boundary contour of the images; according to the boundary point set of the core sample image, identifying core sample end face deformation abnormity and cylindrical surface deformation abnormity; constructing a core sample end face defect identification model, and identifying core sample end face defect types and defect severity; extracting texture features, color features and geometric features of the core sample, and constructing a joint feature matrix; constructing a core sample cylindrical surface defect identification model, and identifying the defect type of the core sample cylindrical surface; performing core sample discarding, finishing and inspection flow distribution according to the deformation and defect identification information of the end surface and the cylindrical surface of the core sample; and determining the compressive strength of the concrete core sample by using an automatic centering and servo loading device. According to the invention, the integrated full-process unmanned closed-loop operation of size detection and strength test is realized, and the requirements of efficiency, consistency and traceability of mass engineering core sample detection are met.

The invention provides an airspace gridding method and device, equipment and a medium, and the method comprises the steps: collecting an airspace boundary geographic data set, taking a horizontal plane at the lowest altitude of a to-be-gridded region as a base surface, determining an airspace boundary based on a boundary contour point geographic coordinate sequence, and carrying out the upward filling of grids based on the base surface, obtaining a gridded three-dimensional airspace and a two-dimensional projection thereof; respectively determining the maximum value point and the minimum value point of the longitude and the latitude to construct a first plane; carrying out gridding on the first plane; and if a plurality of different height values exist, determining the maximum height point, subdividing the airspace between the minimum height point and the maximum height point to obtain second space grids and codes, and screening the second space grids contained in the three-dimensional polygonal region or intersected with the three-dimensional polygonal region. According to the method and the device, the grids and codes of the height dimension between the area bottom plane grids and the boundary contour points are determined, so that the spatial domain is expanded from 2D to 3D grids, and accurate three-dimensional spatial domain gridding is closed.

The application discloses an infrared image segmentation method based on a local region self-organizing mapping algorithm, which comprises the following steps: firstly, a local region self-organizing mapping algorithm is designed, a plurality of feature bias fields are extracted by moving a local sliding window before iterative evolution of a level set, and a plurality of feature local data driving terms are constructed under a multiplicative bias field model framework; secondly, a plurality of feature global data driving terms are calculated, and a plurality of feature hybrid data driving terms are constructed through an adaptive weight function; then, an adaptive regularization function is used to regularize an energy value range of the plurality of feature hybrid data driving terms, and an initial level set is driven to perform iterative evolution; then, in the iterative process of the level set, a hyperbolic tangent function is used to keep the level set in the iterative evolution, and a symbol rule feature of positive outside and negative inside and a distance rule feature with a modulus value of 1 are kept; meanwhile, a mean filter template is used to continuously smooth the level set and eliminate redundant non-boundary contour lines; finally, a gradient descent method is used to continuously iteratively solve a minimum value of an energy function until a convergence criterion of the level set is reached or a maximum iteration number is reached, and then the method is stopped, at this time, a position of a zero level set is output, and image segmentation is completed. The local region self-organizing mapping algorithm can accurately mine a foreground contour from an infrared image, and has good segmentation speed and precision.

The invention discloses an enterprise supplier alternative scheme evaluation method and device based on allowable interval approximate estimation, equipment and a medium, and relates to the technical field of computers, and the method comprises the steps: building an initial decision matrix based on the evaluation value of each supplier alternative scheme of an enterprise under each evaluation attribute, performing approximate estimation by using an allowable interval approximate estimation method and the evaluation values to determine a first tolerance coefficient, and determining an evaluation value reference scheme based on the first tolerance coefficient and the mean value and the standard deviation of the evaluation values; performing category allocation on the supplier alternative schemes by using a target probability classification method, and determining a boundary contour based on an allocation result; and determining a first close degree of the supplier alternative scheme and a second close degree of the boundary contour based on the boundary contour, the initial decision matrix and the evaluation value reference scheme, and evaluating the supplier alternative scheme by using a preset evaluation rule to obtain an evaluation result. Enterprise vendor objects are efficiently classified and rank inversion present in TOPSIS is prevented.

The invention discloses a detail perception enhanced remote sensing small target detection method and system, and the method comprises the following steps: obtaining a to-be-detected image, and carrying out the feature reconstruction of the to-be-detected image, and obtaining a reconstructed feature map; performing wavelet transform down-sampling on the reconstructed feature map to obtain a high-frequency information feature map and a low-frequency information feature map of the reconstructed feature map, and fusing the high-frequency information feature map and the low-frequency information feature map to obtain a fused feature map; processing and fusing the high-frequency information feature map by using a space attention mechanism and a channel attention mechanism to obtain boundary contour features; and taking the fused feature map as the input of a feature pyramid, and combining boundary contour features to obtain a classification and positioning result of the small target. According to the method, boundary contour feature information is extracted by using high-frequency information, accurate extraction of fine-grained boundary contour features is realized in combination with space-channel attention, and finally feature representation of a small target is enhanced through multi-scale feature fusion.

The present invention discloses a periodic and nonperiodic tiling system based on regular hexagons, comprising seven different types of prototiles. Each type of prototiles are based on a regular hexagon structure, and external contours of each type of prototiles are composed of six boundary contours. In a connected pattern enclosed by the six boundary contours, the areas of parts convex from and concave into a regular hexagon are correspondingly equal. A specific shape of an animal, plant or human figure is enclosed by the six boundary contours, and a plurality of specific shapes enclosed by the six boundary contours are engaged with each other to form a tiling surface. The prototiles are provided with a corresponding pattern according to the boundary contours of the specific shape. The present invention is simple in structure and diversified in tiling effect, and has a wide application prospect.

The application discloses an ultrasonic image processing method, device, equipment and storage medium, and the method comprises the steps of detecting a selection instruction for indicating selection of a target object in an ultrasonic image, taking a coordinate indicated by the selection instruction as a starting point coordinate; and performing boundary diffusion identification around the starting point coordinate to identify a boundary contour of the target object from the ultrasonic image. The application takes any point in the target object as the starting point coordinate, and performs boundary diffusion around the starting point coordinate to identify the boundary contour of the target object. The application automatically identifies the boundary of the target object, reduces human intervention, improves the efficiency and accuracy of boundary identification. The boundary contour is identified based on the change of energy coefficients between pixel points, and the change of the energy coefficients conforms to the change of texture features at the internal and boundary parts of the target object, so that the boundary contour can be accurately identified. The whole boundary diffusion process can be superimposed and displayed in the ultrasonic image, and the visualization of the whole boundary diffusion process is realized.

This invention provides a stylizable and deformation-resistant QR code encoding and recognition method based on topological coding, comprising: converting the binary string to be encoded into natural numbers using a bijective function and recursively decomposing it to generate a topological tree structure; determining the spatial allocation ratio based on the tile traces of each node in the topological tree; performing inward shrinking processing on the polygonal region carrying the QR code, recursively dividing it into non-overlapping closed sub-regions according to the ratio, mapping the hierarchical structure of the topological tree to a QR code with a nested structure and stylizing it; scanning the QR code image and performing preprocessing, extracting the boundary contours of the closed regions and constructing a global topological tree, enumerating multiple candidate topological trees and decoding them one by one, and filtering out the valid binary information as the recognition result after CRC verification. This invention achieves the deformation resistance of the QR code through the mapping of the topological tree structure and nested regions, and improves its visual effect and application flexibility through stylization processing.

The present application belongs to the technical field of online early warning, and particularly relates to a kind of necking defect online early warning method of ball tank connecting shell drawing forming necking, comprising the following steps: S1, obtaining a plurality of continuous image sequences in the drawing forming process of ball tank connecting shell, obtaining the three-dimensional geometric model corresponding to ball tank connecting shell;For each frame of image in the continuous image sequence, a plurality of initial necking defect candidate regions represented by boundary contour and initial confidence are obtained using instance segmentation network;S2, the initial confidence of weight, the relevance with the previous frame defect region and the spatial matching degree with the pre-defined high stress area of three-dimensional geometric model are calculated, and the comprehensive risk weight of each initial necking defect candidate region of current frame is obtained, and all initial necking defect candidate regions are arranged in descending order according to comprehensive risk weight.The present application realizes the pre-judgment of defect development, and can issue an alarm when the defect has not yet obviously deteriorated.

The present application relates to a kind of hydrogen leakage safety boundary visualization method and system, belong to new energy technology field.First, the same hydrogen leakage process is carried out using hydrogen leakage experimental device and numerical simulation, data fusion is carried out to the measured hydrogen concentration and simulated hydrogen concentration, and fusion concentration is obtained, then contour curve boundary is randomly taken point until the position of the point is equal to 4%, the position of the point is marked as calibration position, the average gray value is obtained by doing time average processing to the gray value of calibration position, finally, all the pixel points with average gray value in hydrogen leakage gray scale image are marked, and the visual boundary of 4% hydrogen concentration is obtained.The present application uses the strategy of data fusion to improve the accuracy of calibration hydrogen leakage safety boundary profile.

A road center line determination method, device, equipment and storage medium are disclosed. The method comprises: acquiring a first boundary contour line and a second boundary contour line on both sides of a target road; determining an outer bounding rectangle of the boundary contour line and a projection point of each boundary point on the outer bounding rectangle; extending a connecting line between each boundary point and the corresponding projection point, determining a first associated point intersecting the contour line, and taking a midpoint between each boundary point and the corresponding associated point as a feature point corresponding to the boundary point; based on the change direction of the boundary contour line, sorting each first feature point and each second feature point, and based on the sorted feature point sequence, determining a center line corresponding to the target road. Through the technical scheme of the embodiment of the present application, the determination efficiency and accuracy of the road center line can be improved, and the combined center line can perfectly retain the geometric shape of the original two lines.

The invention provides a microneedle patch size determination method and system and a medium, and relates to the technical field of microneedle patch size determination, and the method comprises the steps: obtaining body surface three-dimensional image data of a to-be-applied part, and segmenting the data into a plurality of sub-regions after preprocessing; performing image analysis on each sub-region, extracting local point cloud data, performing curved surface fitting, determining a main curvature coefficient of each sub-region, and further determining a fitting degree index of each sub-region; performing clustering analysis on all the sub-regions according to the fitness indexes, dividing the adjacent sub-regions of which the fitness indexes are located in the same preset interval into the same cluster, extracting local point cloud data of each cluster, performing fitting to generate a region boundary contour line, and incorporating the region boundary contour line into an initial microneedle patch contour line set; and screening and optimizing the initial microneedle patch contour line set to obtain candidate contour lines, smoothing the candidate contour lines to generate a target microneedle patch contour line, and determining an area enclosed by the target microneedle patch contour line as the substrate size of the microneedle patch.

This invention relates to the field of AI analysis technology for medical ultrasound images, specifically to an intelligent ultrasound image analysis method and system based on artificial intelligence. The method includes: acquiring ultrasound image sequences of a target patient and performing standardized preprocessing; extracting deep features through an anatomical structure recognition network to generate an anatomical structure description containing boundary contour coordinate sequences, tissue type classification identifiers, and spatial location depth information; using a lesion feature discrimination model to perform multi-dimensional feature fusion analysis on the three types of information to obtain preliminary lesion localization and attribute descriptions; then verifying spatiotemporal continuity based on adjacent frames of the image sequence to check the consistency of lesion evolution and generate a verification report; comparing the report with a medical knowledge base template to output an auxiliary diagnostic conclusion with diagnostic opinions and confidence ratings. This invention can enrich the dimensions of anatomical structure representation, improve the accuracy of lesion discrimination, ensure the temporal consistency of analysis results, and enhance the reliability of auxiliary diagnostic conclusions.

The present disclosure relates to a method and a device for image instance segmentation, as well as a device and a product. The method comprises: determining an interest region feature of a target object and a target region image of the target object in an input image based on the input image. The method further comprises: extracting a shape feature of the target object based on the target region image. The method further comprises: determining a fused feature of the target object based on the interest region feature and the shape feature. In addition, the method further comprises: performing instance segmentation of the input image based on the fused feature to generate a mask instance of the target object in the input image. In this way, the shape property of the target object is taken into account in the fused feature of the target object.In the embodiments described in this disclosure, image instance segmentation is performed by using the fused feature of the target object, thereby enabling the focus to be placed on the entire region of the target object, particularly regions obscured by other target objects. This, in turn, allows for more accurate detection of the boundary contour of the object region.

The application discloses an image processing method, and relates to the fields of auxiliary driving and image processing. The method comprises the following steps: for each sub-block image, acquiring the reference feature pixel point quantity, the first feature pixel point and the second feature pixel point of the to-be-fused boundary contour of the sub-block image; acquiring the coordinate data of the reference feature pixel point of the to-be-fused boundary contour according to the above parameters; determining the fusion boundary of the sub-block image according to the coordinate data corresponding to the first feature pixel point, the second feature pixel point and the reference feature pixel point; and performing image fusion on each sub-block image according to the position relationship between each fusion boundary and each sub-block image. The technical scheme of the embodiment of the application not only realizes the accurate construction of the fusion boundary conforming to the actual contour of the sub-block image, but also avoids the problems of obvious splitting, ghosting or display fault at the splicing position of the sub-block image, thereby improving the content continuity, integrity and clarity of the fused image.

The application discloses a parametric design method, device, equipment and medium for a sub-instrument panel armrest section, which comprises the following steps: firstly, based on a unified design reference system, a standardized parametric digital model of a full-level feature contour of an embedded armrest, adjustable size parameter variables and a linkage correlation relationship among the three is constructed to form a reusable design basis; secondly, target project design input is acquired and a boundary contour is extracted, and the model is called to complete reference matching and contour correlation; finally, by adjusting the size parameters, the feature contour is automatically and synchronously updated by means of the linkage relationship, and the complete section drawing process does not need to be repeated, so that efficient parametric design of the armrest section can be completed.

The application provides a high-speed motor operation data intelligent analysis method and system, and relates to the technical field of data processing.The method comprises the following steps: performing a preprocessing operation on basic operation information and surrounding working condition factors to form initial integrated information; performing boundary contour calculation on a multidimensional data point set for the initial integrated information, and extracting an outer convex boundary point set of data distribution; determining a coverage range of data distribution based on the outer convex boundary point set, and establishing a minimum convex polygon verification interval containing all data points; performing adaptive hierarchical cutting on the verification interval according to a data density gradient and a working condition change amplitude to form a plurality of cutting sub-regions with different characteristic densities; extracting geometric features, distribution densities and boundary curvatures of each cutting sub-region as characteristic parameters, and mapping and converting the characteristic parameters into stability evaluation parameters and abnormality evaluation parameters of corresponding verification points.The application realizes accurate evaluation, real-time abnormality identification and dynamic adaptation of the running state of a high-speed motor.

The application discloses a ground stair identification and stimulation method and system for a bionic vision system, and the method comprises the following steps: acquiring depth information and spatial reference information of a target scene; performing constraint on the depth information based on the spatial reference information to identify and extract a passing area plane in the scene and determine a boundary contour of the passing area; performing hierarchical structure identification on the depth information of a non-passing area to extract a stair structure with a hierarchical relationship; performing adaptive abstraction processing on the boundary contour of the passing area and the stair structure based on the number of electrode channels of the bionic vision system to generate a sparse dot array expression matched with the number of electrode channels; and mapping the sparse dot array expression to an electrode array coordinate system of the bionic vision system to generate an adaptive stimulation mode and outputting the same. The safety hazards of misjudging a path, missing a stair or failing to perceive a sudden structure of a blind user in a complex terrain due to insufficient information expression are effectively solved.

The present application relates to the technical field of defect detection, in particular to an electronic component defect visual detection system and method, the system comprising an image illumination control module, a defect boundary identification module, a confidence level calculation module, a result level classification module and a structure repetition elimination module.In the present application, the edge gray value and the inside pixel gray value of the packaging pin and the soldering point area are extracted, and the brightness gradient change trend is calculated, to realize dynamic adaptation of image illumination control, improve the clear identification of defect contour, build a defect intensity index sequence, and combine historical fluctuation interval difference to complete dynamic adjustment of the confidence interval, so that the defect level evaluation has stable distinguishing ability, compare the defect intensity value with the upper and lower limits of the confidence interval, combine the label attribution logic to form a clear distribution layer, match the boundary contour with the gray texture feature to realize structure repetition elimination between defects of the same level, and improve the independence and annotation accuracy of the detection layer.

This invention discloses a blind zone detection method for handheld point cloud scanning based on the interaction between the acquisition area and the boundary, relating to the fields of 3D scanning and point cloud data processing technology. The method first acquires the device pose and local point cloud, preprocesses it, and then transfers it to a global horizontal coordinate system. A boundary feature extraction engine is used to identify room boundary lines and project them to form a 2D boundary contour map. Simultaneously, an effective acquisition loop is dynamically constructed and projected based on the effective measurement range parameters of the laser scanner and the current ambient lighting conditions. Boolean operations are performed on the effective acquisition loop and the boundary contour map to identify uncovered boundary segments as potential blind zones. The occlusion state is determined using the temporal information of continuous frame point clouds and a ray casting algorithm to identify the effective blind zone. This invention reduces the 3D integrity determination to 2D planar geometric intersection, eliminating the need for prior models and enabling real-time identification of dynamic occlusion and acquisition blind zones. This results in high computational efficiency and significantly improves the success rate and coverage integrity of handheld scanning operations.

Provided in the present disclosure are an obstacle avoidance processing method for a cleaning device, and a cleaning device, a cleaning system, a computer-readable storage medium and an electronic device. The obstacle avoidance processing method for a cleaning device comprises: detecting the surrounding environment of a cleaning device by means of an obstacle sensor, so as to generate a low-obstacle map corresponding to the surrounding environment, wherein the low-obstacle map is at least used for indicating a boundary contour of a low obstacle in the surrounding environment; on the basis of the boundary contour of the low obstacle in the surrounding environment, determining an obstacle avoidance path for the low obstacle; and when the cleaning device approaches to the low obstacle, the cleaning device moving along the obstacle avoidance path.

The present application relates to the field of industrial vision technology, and proposes an online early warning method and system for edge cracks of cold-rolled strip steel based on industrial vision, which comprises the following steps: taking the connected boundary contour, geometric center position and maximum extension length along the running direction of the strip steel in the differentiated crack feature area of the edge surface of the strip steel as the crack labeling data; determining the crack evolution trajectory of the edge surface of the strip steel according to the geometric center position of the same crack object at different times and the maximum extension length at different times in the crack labeling data; and dynamically comparing the current crack length growth rate in the crack evolution with the preset crack propagation reference state to obtain the early warning information of the edge surface of the strip steel; and the present application can improve the efficiency of online early warning for edge cracks of cold-rolled strip steel.

The invention relates to the technical field of coal seam image processing, and particularly discloses a coal seam separation boundary extraction method based on deep learning, which is used for solving the problem that pseudo-edges, fractures and horizon drifts easily occur in interface segmentation and edge extraction due to interference of water film fogging, reflective distortion and texture gradual change on a composite roof roadway drill hole wall image. Comprising the steps of hole wall image preprocessing and cylindrical surface expansion correction, near-far contrast difference calculation and equivalent driving quantity thermodynamic diagram generation, imaging quality and contour continuity weight construction, multi-source feature fusion depth segmentation and boundary contour extraction, contour and mask write-back weight self-updating and closed-loop iteration convergence control. According to the method, through near-far side comparison driving thermodynamic diagrams, imaging quality and contour continuity weight constraints and in combination with depth segmentation closed-loop iteration, the boundary contour of the potential structural plane of the separation layer is stably extracted, and false edges, fractures and horizon drifts caused by fogging water films and crack drill marks are reduced.