19 results about "Contour set" patented technology

The application relates to a region detection method and device, computer equipment and a computer readable storage medium. The method comprises the following steps: obtaining a layer design diagram of a forming layer of a circuit board; the circuit board comprises a plurality of circuit board blocks, the forming layer is a layer comprising a block contour of each circuit board block, and the block contour comprises an inner contour and an outer contour of the circuit board block; contours in the layer design diagram of the forming layer are extracted to obtain a contour set; the contours in the contour set correspond to associated contour information, the associated contour information of the contour is used for representing sub-contour information and parent contour information of the contour; contours with associated contour information satisfying preset associated contour information are selected from the contour set to obtain inner contours corresponding to each circuit board block respectively, and the regions of the circuit board blocks are determined based on the inner contours corresponding to each circuit board block respectively. The method can improve the efficiency of determining the regions of the circuit board blocks.

The application provides a method for judging the feasibility of inclined three-dimensional printing of a shell model, comprising: slicing an STL model to be judged according to a given inclination angle, placing the inclined slice horizontally, and obtaining the contour data of the horizontal slice; respectively biasing each contour in each slice layer inward and outward, and performing Boolean intersection operation to obtain the multi-fork tree structure contour set of each contour; and judging whether the contour in each slice layer can be stacked in a certain connected region in the multi-fork tree structure contour set of the slice layer below, and further judging whether the model can be printed in an inclined manner. The judgment method of the application is aimed at the printing structure of a shell mold, can quickly judge the feasibility of inclined printing according to the mold slice information, can assist the operator of the printer to set the model placement posture, select a suitable slice angle, and reduce the printing trial and error cost; and the judgment method of the application has wide applicability, high efficiency and strong stability.

The application relates to a high-precision map double-yellow-line post-processing method and device and electronic equipment. The method comprises the following steps: obtaining a semantic segmentation result of high-precision map data, and extracting an initial contour set from the semantic segmentation result; the initial contour set comprises candidate contours corresponding to double yellow lines; the shortest Euclidean distances between all candidate contours in the initial contour set are calculated, and according to the distance calculation result and a preset adhesion threshold, a pair of adhesion contours is determined from the candidate contours; for each contour in the pair of adhesion contours, the corresponding two contours in the pair of adhesion contours are translated by a preset translation amount in the normal direction of the contour, and a separated contour set is obtained; and a double-yellow-line geometric element is generated according to the separated contour set. The scheme provided by the application can effectively solve the misrecognition and missed detection problems of double yellow lines caused by pixel adhesion in complex scenes such as bad weather, improve the double yellow line recognition accuracy and data integrity, and further improve the production accuracy of the high-precision map.

The present application relates to the field of image and video compression technology, and more particularly to a semantic contour generation type image coding method based on key point optimization, comprising obtaining an image to be processed and inputting a visual language model and an instance segmentation model respectively to generate an overall text description and a complete contour mask; inputting the complete contour mask into a VQ-VAE encoder to output a codebook index sequence set; for the complete contour mask, a differentiable model based on Gumbel-Softmax is used to adaptively select a group of key points; a VQ-VAE decoder generates a preliminary reconstructed contour according to the codebook index sequence set; a lightweight optimization network performs geometric correction and refinement on the preliminary reconstructed contour based on the key points to obtain an optimized contour; and the optimized contour and the overall text description are taken as control conditions to generate a high-quality reconstructed image through a conditional generation model; and the present application replaces a non-structured global pixel edge map with a structured contour set to eliminate redundant edge information in the background and the object interior.

The invention discloses a workpiece defect detection method and system based on sub-pixel features, and relates to the technical field of defect detection.The workpiece defect detection method comprises the steps that a machine vision device is used for collecting a workpiece image, an edge contour set of a workpiece is obtained, and the workpiece is divided into edge lines of different categories; the method comprises the following steps: performing data set division on historical workpiece defect data according to different edge line categories, performing image segmentation and morphological processing on defect samples in each historical workpiece defect data set, obtaining defect areas, performing sub-pixel-level edge extraction, and analyzing symbolic defect sub-pixel features in defect edges; after a to-be-detected workpiece image is preprocessed, a region of interest is selected for sub-pixel-level edge extraction, and the similarity between the sub-pixel-level edge of the region of interest and the sub-pixel features of the symbolic defect is calculated; and judging defect information of the workpiece according to the similarity. According to the invention, the edge is accurately positioned through sub-pixel-level edge detection, the detection precision is improved, and high-precision rapid detection of workpiece quality is realized.

The embodiment of the invention provides a method for determining an adjacency relation between spatial contours, a method and device for determining a spatial adjacency matrix, a computer readable storage medium and electronic equipment, and relates to the technical field of data processing. The method comprises the steps that a target contour set corresponding to the ith contour is obtained, a first relation and / or a second relation exists between the jth contour and the ith contour in the target contour set, the first relation is that the two contours have a shared boundary on the boundary and the internal spaces are not overlapped, and the second relation is that the two contours are partially overlapped on the internal spaces; obtaining an intersection between the boundary of the i-th contour and the j-th contour; obtaining objects of which the types are line segments in the intersection, and obtaining a target boundary; and if the length of the target boundary meets a preset condition, determining that the two space contours are adjacent to each other. According to the embodiment of the invention, the judgment accuracy of the adjacent state between the contours can be improved, and the generation of a more reasonable space adjacent matrix with more practical application significance is facilitated.

The application provides a 3D printing data processing method, system, device and medium of a porous implant. The method comprises: slicing a three-dimensional model of the porous implant to generate an initial two-dimensional cross section; extracting a solid contour in the cross section and calculating a solid area ratio and a cross section shape complexity parameter thereof; based on the parameters, classifying the solid contour into a first contour representing a junction and a second contour representing a strut; based on the distribution state of the first contour in each cross section, assigning a target slice layer thickness to a corresponding height interval and performing re-slicing; simplifying the first contour and the second contour in the re-sliced cross section respectively, and the simplification accuracy of the first contour is higher than that of the second contour; and generating printing data based on the simplified contour set. The application can reduce the consumption of computing resources and ensure the mechanical forming strength at the junction.

The application discloses a k representative G-Skyline query method, comprising the following steps: S1, preprocessing a whole data set D to be queried to obtain an ordered contour set SKY(D)={s1, s2,..., s n}; S2, constructing a dominance relation graph according to the whole data set D to be queried; S3, constructing a dominance order graph based on a replacement strategy and an expansion strategy; S4, pruning a set of data tuples in the dominance order graph which cannot become a k representative G-Skyline query result according to a pruning strategy; and S5, returning a set of the first k data tuples with the largest dominance area in the pruned dominance order graph as a final result, so that the problem of high calculation cost in the existing k representative G-Skyline query method is solved.

Embodiments of the present disclosure provide a monocular vision-based robot obstacle avoidance method, system, device and medium, comprising: outputting a category label of a target obstacle through a deep learning model for target detection on a target image taken for a target scene where a robot is located; performing contour extraction on the target image to obtain a closed contour set of the target obstacle; in a target region of interest, filtering the closed contour set based on a geometric constraint criterion to obtain a plurality of target feature samples; performing geometric ranging calculation according to the category label and a sample average pixel area of the plurality of target feature samples to obtain a real-time distance between the robot and the target obstacle; when the real-time distance is less than or equal to a safety threshold, comparing an area difference rate between a preset central region of interest and a side region of interest with a difference rate threshold, determining a planar spatial distribution trend of the target obstacle and determining a target obstacle avoidance strategy. The reliability of robot obstacle avoidance can be improved.

The invention discloses a piston top valve pit detection method and device based on image recognition, and particularly relates to the technical field of valve pit detection. The method comprises the following steps: acquiring a multi-angle polarization image sequence at the top of a piston, and carrying out lens distortion correction and gray scale unification processing to form a standard polarization image set; calculating a polarization degree image and a polarization angle image pixel by pixel based on the standard polarization image set, and fusing to generate a polarization reflection feature map; segmenting a sediment area at the top of the piston by adopting a polarization consistency constraint method based on the polarization reflection characteristic pattern, and outputting a sediment mask pattern; extracting candidate boundaries of the valve pit at the top of the piston by adopting an edge curvature continuity constraint method based on the polarization reflection feature map and outputting a candidate contour set; and carrying out space overlap suppression and contour continuity completion processing on the sediment mask pattern and the candidate contour set to obtain a real contour of the valve pit without sediment interference, calculating position coordinate parameters and size parameters of the valve pit, and generating a detection result of the valve pit at the top of the piston.

The invention relates to the technical field of spatial data analysis, in particular to a contour line drawing method and device, electronic equipment and a storage medium, and the method comprises the steps: carrying out the spatial smoothing of input two-dimensional lattice point field data through employing a Gaussian separable convolution mode based on a set smoothing scale parameter, and obtaining a smoothed lattice point field; and based on the smoothed grid point field data, generating a contour line set through contour line tracking and spatial topology analysis. According to the method, high-efficiency smoothness is realized through Gaussian separable convolution, the innermost contour line is automatically identified based on topology analysis, intelligent labeling of high-voltage and low-voltage centers is realized in combination with regionalized extreme value extraction and spatial overlay analysis, and the data is greatly compressed while the precision is ensured by adopting third-order Bezier curve fitting, so that the accuracy of the high-voltage and low-voltage centers is ensured. And finally, the contour line with smooth vision and clear structure is output, and the visualization quality and the automation level are improved.

The invention discloses a Raman temperature and humidity dynamic detection method based on self-adaptive light intensity adjustment, and belongs to the technical field of Raman detection. The method comprises the steps of collecting and processing spectral data to generate a preorder feature contour set, constructing a temperature and humidity parameter detection pair, simulating interference to generate a postorder feature contour set, analyzing contour feature parameters, calculating fluctuation, screening an optimal detection pair and feeding back. The system correspondingly comprises a spectral data acquisition and processing module, a temperature and humidity parameter configuration module, a feature contour analysis module, a fluctuation calculation and screening module and a feedback execution module. By dynamically adapting the optimal temperature and humidity parameter detection pair, the problem of insufficient detection precision caused by light intensity interference and low parameter matching degree in traditional Raman temperature and humidity detection is solved, manual intervention on parameter adjustment is not needed, the detection stability and efficiency are improved, and the method is suitable for various high-precision temperature and humidity detection scenes such as industrial production and environment monitoring.

A computing device configured to receive at least a radiological image, encode, using an encoding module, a low-dimensional image as a function of the radiological image, wherein the encoding module is configured to preprocess the radiological image, generate a segmented representation of the radiological image by applying edge detection to the preprocessed radiological image, generate a reduced contour set by performing contour simplification and dimensionality reduction on the segmented representation, and generate a color-coded contour map by coloring coding one or more contours of the reduced contour set, and input the low-dimensional image into at least a machine learning process as training data.

The invention provides a plastic package sorting method and system, electronic equipment and a storage medium, and the method comprises the steps: carrying out the spectral feature analysis of hyperspectral data, and determining the material type of plastic; performing morphological analysis on the image model data, performing multi-modal matching verification on a shape analysis result and spatial distribution of the hyperspectral data, and determining geometric attributes; generating a correlation model based on the specified sorting strategy database, the material type and the geometric attribute; extracting a pixel-level contour set of the object according to the image model data and mapping the pixel-level contour set to corresponding jet equipment parameters to obtain a target pore matrix; and determining an air valve control instruction of the air injection equipment according to the air injection equipment parameters, the correlation model, the target air hole matrix and the conveying speed. Through double combination of hyperspectral data and image model data, the problem that the defect that homochromatic heterogeneous plastics (such as transparent PET and PVC) cannot be distinguished by a single vision technology is solved.

The invention relates to the technical field of semiconductor detection, discloses a wafer level packaging structure surface defect detection method, and particularly relates to a wafer through hole stacking quality detection method and system and a readable storage medium. According to the embodiment of the invention, through hole segmentation is realized through a combined segmentation method, and the combined segmentation method comprises a low-cost and low-precision traditional segmentation method and a high-precision segmentation method based on a segmentation model. According to the embodiment of the invention, the image is initially segmented through a traditional segmentation method to determine whether the current segmentation result meets the requirement or not, and if the current segmentation result does not meet the requirement, the segmentation model is adopted for accurate segmentation to obtain a first through hole and a second through hole in the wafer and a contour set corresponding to the two through holes. And determining corresponding geometric features based on the contour sets corresponding to the two, and finally determining whether the quality of the through hole meets process requirements based on the geometric features.

The application discloses a post-processing method and system for converting natural resource element mask grid data into high-quality vector polygons, and the method comprises the following steps: obtaining natural resource element mask grid data through a deep learning model; pre-processing the natural resource element mask grid data, eliminating redundant elements, and obtaining a preliminary contour set; performing coordinate conversion on the preliminary contours in the preliminary contour set, and performing range contour simplification on the preliminary contours by using an improved Douglas-Peucker algorithm to obtain a simplified contour set; performing boundary smoothing processing on the simplified contours in the simplified contour set by using Gaussian smoothing to obtain an optimized contour set; performing vector topological compliance checking on the optimized contours in the optimized contour set, then converting the checked and repaired contours into GeoDataFrame, and saving the GeoDataFrame as an SHP format vector file, and completing the post-processing.

This invention discloses an intelligent detection system and method for automotive headlight intensity, comprising the following steps: S1, acquiring headlight illumination data through an intelligent sensor; S2, constructing a spatial mapping matrix, denoising, linearizing, and back-projecting the illumination data to generate an illumination matrix; S3, performing intensity segmentation on the illumination matrix, extracting light spot regions, and constructing a contour set; S4, constructing a light pattern description tensor using the illumination matrix and the contour set, and generating a feature vector set; S5, matching the feature vectors with a benchmark library to generate a judgment matrix and fusing and calculating the output light intensity, light spot offset, and cutoff line anomaly results; S6, generating a detection report and writing it to a storage path to form a detection summary. This invention achieves high-precision intelligent detection of automotive headlight intensity and light spot anomalies through intelligent acquisition and multi-dimensional feature judgment, improving detection accuracy and automation.

The application provides a model printing method based on a 3D printing device, an electronic device and a storage medium. The method comprises the following steps: acquiring a shape parameter of a to-be-printed model, and acquiring a contour set of the to-be-printed model according to the shape parameter; receiving a cutting line characteristic parameter, and generating a cutting surface of the to-be-printed model according to the cutting line characteristic parameter; dividing the to-be-printed model into a plurality of to-be-printed sub-models according to the cutting surface and the contour set; receiving a model printing sequence, and printing the plurality of to-be-printed sub-models in sequence according to the model printing sequence. The application can reduce the printing cost of a 3D printing device when printing a large model, and can realize multi-color printing of the model.

The embodiment of the invention provides a high-reflection groove recognition method and device based on a 3D camera. The method comprises the steps that a two-dimensional contour set of a groove of a workpiece is obtained through a 3D scanning camera; supplementing missing segments of the two-dimensional point sequence based on the continuity of the two-dimensional point sequence and the groove model feature library; selecting a candidate feature point set based on a multi-scale chord-to-point distance accumulation algorithm; and filtering the candidate feature point set based on set verification information, and determining the three-dimensional features of the groove according to the candidate feature points obtained after filtering and the attitude information of the 3D scanning camera. Through point cloud intelligent completion based on continuity and structure prior, multi-scale robust curvature feature fusion extraction and multi-stage prior knowledge verification and filtering, high-precision, high-robustness and high-adaptability automatic identification of key geometric feature points of the groove under the conditions of high reflection, strong noise and local serious data loss is realized. And stable and reliable visual guidance is provided for subsequent robot automatic welding.