51 results about "Plane segmentation" patented technology

The invention relates to the technical field of unmanned aerial vehicle navigation, and discloses an unmanned aerial vehicle navigation method and system based on computer vision, and the unmanned aerial vehicle navigation method based on computer vision comprises the following steps: collecting a current visual field image of an airborne camera of an unmanned aerial vehicle, and carrying out adaptive image enhancement; performing scene semantic understanding; semantic perception feature extraction is carried out; feature matching and outer point elimination are carried out; floor identification is carried out, and multi-sensor fusion positioning is carried out; plane segmentation, structural constraint extraction and map normalization are carried out; performing visual navigation evaluation, judging a navigation mode according to an evaluation result, and generating an exploration control instruction quantity based on the navigation mode; generating a cross-floor path sequence, optimizing a track in real time, and performing control fusion in combination with a navigation mode identifier and an exploration control vector; according to the method, the navigation problem of the unmanned aerial vehicle in complex scenes such as smoke interference, multi-story buildings and illumination dramatic change in a GPS denial environment is solved, and high-precision autonomous navigation is realized.

The invention discloses a deformed workpiece twin modeling method based on measurement point cloud and three-dimensional CAD model matching. The deformed workpiece twin modeling method comprises the steps that an improved GFO-ICP registration algorithm is adopted to solve the initial coarse registration problem of the measurement point cloud and a CAD model in combination with geometric feature constraint and self-adaptive step length optimization; extracting a key geometric curved surface of the point cloud by adopting a clustering-based plane segmentation algorithm, and extracting a semantic feature vector of a segmented region in combination with a Pointnet + + network; a cosine similarity matching mechanism is constructed by integrating mixed descriptors of geometric features (areas and normal vectors) and semantic features, and accurate binding of a point cloud curved surface and a curved surface of a CAD model is achieved. According to the method, the problem of assembly geometric deviation caused by workpiece deformation due to manufacturing errors in compressor and fan rotor assembly is solved, multi-link error accumulation is effectively processed, digital twin modeling of high-precision deformed workpieces is achieved, and the assembly simulation effect and intelligent manufacturing application are improved.

The invention relates to a robot dog dynamic balance and fall self-recovery method and device for a complex terrain, and the method comprises the steps: building a robust local elevation map through fusing multi-sensor data and quantifying the motion uncertainty of a robot; and a safe foot region set defined by the convex polygon is extracted through geometric feature analysis and plane segmentation. Then, collaborative planning of the trunk and the landing point is carried out in combination with a user instruction and the region set. In a control layer, a nonlinear model predictive control algorithm is adopted, a safety area is used as a hard constraint, and incomplete speed constraint is applied to a wheel type foot end, so that an optimized trunk track and a foot end force reference are output. The whole-body controller converts an upper-layer instruction into an accurate joint torque through a multi-task optimization framework with a weight. When instability is detected, the system can autonomously plan and execute a recovery action sequence including curling, controllable rolling and cooperative supporting.

The invention provides a weld joint extraction method based on multi-dimensional information fusion. The method comprises the following steps: S1, collecting a Texture image and depth data of a weld joint; s2, inputting the Texture image into an image recognition model to obtain a mask image of the welding seam; s3, binarization processing is carried out on the mask image, and pixel coordinates of the welding seam are extracted; s4, aligning the pixel coordinates with the depth data to generate an initial welding seam point cloud; s5, the initial welding seam point cloud is processed through plane segmentation and a fitting algorithm, and welding seam feature points are extracted; and S6, obtaining an accurate welding seam point cloud by using a straight line fitting algorithm or a plane intersection fitting algorithm. By combining the two-dimensional image segmentation and the three-dimensional point cloud processing technology, the problem that accurate three-dimensional information extraction and background noise interference resistance of the weld joint are difficult to consider simultaneously in a complex industrial scene by a traditional method is solved, the precision and robustness of weld joint extraction are improved, the processing efficiency is ensured, and the method is suitable for popularization and application. And a high-reliability three-dimensional path planning basis is provided for an automatic welding system.

The invention discloses a sugarcane live-action three-dimensional phenotype extraction method based on a 3D Gaussian splashing technology. The method comprises the following steps of data acquisition, wherein a 360-degree omnibearing shooting mode is executed around a single sugarcane plant for shooting or video recording to obtain complete image or video data of the sugarcane plant; preprocessing the data; 3D model construction comprises the following sub-steps: SfM sparse reconstruction: instance segmentation; the method comprises the following steps: carrying out 3D Gaussian Splicing guided by a mask; point cloud data processing includes the following sub-steps: point cloud preprocessing including voxel downsampling, statistical filtering and radius filtering; segmenting and identifying stalks / leaves; the segmentation comprises RANSAC plane segmentation, DBSCAN clustering and region growth; the phenotype analysis comprises grid reconstruction and area / volume / perimeter / plant height calculation. According to the invention, the problems of difficult data acquisition, large workload and incomplete information acquisition in the prior art are solved.

The invention provides a turbine blade wear-resistant layer welding path planning method based on point cloud processing. The method comprises the steps that firstly, a structured light camera and a welding robot tail end coordinate system and a conversion matrix of the welding robot tail end and a world coordinate system are determined through calibration; scanning the turbine blade by using a structured light camera to obtain original three-dimensional point cloud data of the turbine blade, and recording coordinates of a welding gun under a tail end coordinate system of a welding robot; performing point cloud filtering, Euclidean clustering and plane segmentation on the original three-dimensional point cloud data to obtain point cloud data of a to-be-welded surface; the point cloud of the surface to be welded is cut through a principal component analysis algorithm to obtain a preliminary path, a complete path can be obtained through straight line fitting and path logic sorting processing, and then coordinates of a welding gun in a coordinate system at the tail end of the welding robot and the complete path are substituted into a coordinate transformation matrix; coordinates of all the path points in the world coordinate system can be obtained, so that path planning is completed. According to the method, automatic path planning can be realized, and production efficiency is improved.

This invention relates to a method for inverting remote sensing data from overlapping strong and weak absorption lines of greenhouse gases, belonging to the field of satellite remote sensing technology. Specifically, it includes the following steps: Step 1: Constructing an optical system, including a satellite platform, camera 1, and camera 2; camera 1 and camera 2 are mounted at the bottom of the satellite platform; camera 1 and camera 2 are mirrored; the imaging spectral band of camera 1 moves from the weak absorption spectral band to the strong absorption spectral band along the positive push-broom direction to the negative push-broom direction; the imaging spectral band of camera 2 is in the opposite direction; the optical system moves horizontally from left to right to achieve horizontal push-brooming of greenhouse gases on the ground, obtaining a dual-camera fused image; the dual-camera differential absorption spectroscopy algorithm is used to invert the dual-camera fused image to obtain the greenhouse gas column concentration; this invention, based on the spectral characteristics of strong and weak absorption lines of greenhouse gases, adopts a detector focal plane segmentation method, and the dual cameras achieve differential absorption and inversion in the same region, which can guide the design of greenhouse gas detection micro / nano satellite payloads.

The present invention relates to the technical field of three-dimensional plane projection. Specifically disclosed are a plane projection method and plane projection apparatus on a surface of a three-dimensional object. The method comprises: determining a projection distance of a projection apparatus; determining a projectable range on the basis of the projection distance; performing plane segmentation on a three-dimensional object for projection, so as to obtain a plurality of sub-planes; on the basis of the sub-planes and the projectable range, determining an arrangement mode of the projection apparatus; and arranging a plurality of projection apparatuses on the basis of the arrangement mode, and executing a screen projection operation on said three-dimensional object by means of the plurality of projection apparatuses. By improving a conventional screen projection method, a surface of a three-dimensional object is segmented, and projection is separately performed on different sub-planes, thereby ensuring sufficient projection clarity and satisfying actual requirements.

The invention provides a binocular image-based parallax plane segmentation method and device, and the method comprises the steps: obtaining an original image and camera internal parameters, the original image comprising left and right views, a depth image and a normal vector image; generating a plane segmentation image and plane parameters according to the original image, and obtaining a true value of plane parallax according to the camera internal parameters; the plane segmentation image, the plane parameter and the true value of the plane parallax are input into a pre-trained binocular parallax plane segmentation network to obtain a plane segmentation result output by the binocular parallax plane segmentation network, and the plane segmentation result comprises a predicted plane segmentation image, a plane parameter image, a parallax image and a three-dimensional reconstruction result; wherein the binocular parallax plane segmentation network is obtained by training a pre-constructed deep neural network by using an image sample. The position and direction information of the plane is fully utilized, the reconstruction precision of the shielding area is improved, and the three-dimensional reconstruction effect is improved.

The invention relates to the technical field of ship plane segmentation, and discloses a plane segmentation assembly line manufacturing method which comprises the following steps: step 1, main board splicing operation; step 2, SAW welding and fixing; step 3, carrying out segmented turning-over operation; step 4, configuring and scribing a longitudinal bone; 5, longitudinal installation operation is carried out; step 6, longitudinal welding operation; 7, conveying to a jig frame section by section; 8, pulling-in operation of the inner structural material is carried out; ninthly, inner structural material assembling work is conducted; 10, welding operation is conducted on the inner structural material; step 11, checking and repairing operation; step 12, pre-outfitting operation is carried out; and step 13, carrying out subsection transportation operation. The flow from one station to the next station is segmented according to the fixed rhythm, so that the waiting time in the traditional fixed station production is eliminated, and the production process is smooth; the labor investment is effectively reduced; the conveying efficiency is effectively improved through the conveying trolley and the plate chain conveying device, conveying is carried out according to a specific rail, the problem of conveying errors is avoided, and the conveying accuracy is improved.

This invention provides a method for training a curved surface intelligent spraying model, a spraying method, and a system thereof, relating to the field of robot path planning technology. The curved surface intelligent spraying method includes: acquiring point cloud data of an object using a depth camera; performing voxel filtering, statistical filtering, and RANSAC plane segmentation preprocessing on the point cloud data; calculating the centroid and PCA principal component orientation of the preprocessed point cloud; rasterizing the space containing the point cloud according to a preset grid size and number of rows and columns to generate a grid map; using a DQN model for path planning; and having the robot perform spraying operations according to the planned path. This invention also provides a method for training a curved surface intelligent spraying model and a curved surface intelligent spraying system to achieve optimal path coverage of all sprayable areas, avoiding unnecessary movement.

The invention discloses a modular room size automatic measurement method based on a structure chart and a measurement chain, and belongs to the crossing field of computer vision and artificial intelligence. The method comprises the following steps: firstly, acquiring three-dimensional point cloud data of a concrete module unit to be measured by using a high-precision laser scanner, and obtaining a structure plane set with millimeter-level precision through structure priori guided graph-cut plane segmentation; semantic classification is carried out by extracting multi-scale geometric features and room-level context features, and a plane set with labels is output; constructing a structure chart based on the spatial topological relation and the semantic information; then defining a basic measurement unit, and automatically compiling a high-level measurement task into an ordered measurement chain; and finally, sequentially executing the measurement chain, binding real point cloud coordinates, and outputting a measurement result with deviation judgment. According to the method, full automation from the point cloud to the structure diagram to interpretable measurement is realized, and the method has millimeter-level precision, interpretability and module reusability, and is suitable for factory quality inspection of concrete modular buildings.

The application discloses a point cloud filtering method and system for a laser scanner, and the point cloud filtering method comprises the following steps: acquiring point cloud data to be processed; generating a two-dimensional image according to the spherical coordinates of all data points in the point cloud data, wherein the horizontal coordinate of each pixel of the two-dimensional image corresponds to the azimuth of a data point, the vertical coordinate corresponds to the elevation of the data point, and the pixel value of the pixel corresponds to the depth value of the data point; for each data point, the size of the filtering core of the data point is acquired according to the depth value; for a target data point, the neighbor data points of the target data point are found on the two-dimensional image according to the size of the filtering core; the depth value of the target data point is acquired by using the neighbor data points, and the coordinates of the target data point are updated by using the filtered data. The application greatly shortens the operation time of point cloud processing, obtains point cloud data with less noise, thinness and accuracy in a shorter operation time, and improves the accuracy of point cloud measurement and the quality of plane segmentation.

The application provides a welding seam extraction method based on multi-dimensional information fusion, comprising the following steps: S1, collecting a Texture image and depth data of a welding seam; S2, inputting the Texture image into an image recognition model to obtain a mask image of the welding seam; S3, performing binaryzation processing on the mask image to extract pixel coordinates of the welding seam; S4, aligning the pixel coordinates with the depth data to generate an initial welding seam point cloud; S5, processing the initial welding seam point cloud by using a plane segmentation and fitting algorithm to extract characteristic points of the welding seam; and S6, obtaining an accurate welding seam point cloud by using a straight line fitting algorithm or a plane intersection fitting algorithm. By combining two-dimensional image segmentation and three-dimensional point cloud processing technology, the application solves the problem that a traditional method is difficult to simultaneously consider accurate three-dimensional information extraction and anti-background noise interference of the welding seam in a complex industrial scene, improves the precision and robustness of welding seam extraction, and guarantees the processing efficiency, thereby providing a high-reliability three-dimensional path planning basis for an automatic welding system.

Even at locations where the point cloud shape of the edge that represents a cross-sectional change cannot be obtained, the completed value of the slope protection structure can be accurately calculated. [Solution] The system includes a plane extraction means 105 that extracts a plane based on the normal vector at each point in the 3D point cloud data, a plane separation means 106 that separates the plane into a slope frame portion and a bottom portion, a center point calculation means 107 that clusters the point cloud data of the bottom portion and calculates a center point for each clustered bottom portion, a selection means 108 that generates cross-sectional clusters for the point cloud data included in the cross-sections that traverse vertically and horizontally from the center point, and selects the cross-sectional slope frame portion and the cross-sectional slope bottom portion based on the average elevation of the cross-sectional clusters, an extraction means 109 that extracts the cross-sectional slope frame portions near the left and right sides of the cross-sectional slope bottom portion that includes the center point, and a completed value calculation means 110 that calculates the completed value based on the extracted cross-sectional slope frame portions.

Disclosed is a method and an apparatus for plane segmentation of a three-dimensional rendering gradient multi-level plane. The method includes: performing a grayscale conversion on a target image, and performing a projection and pixel grouping on a first image processing result obtained by the grayscale conversion, so as to obtain a second image processing result including a plurality of triangular patches; performing determination of vertexes for each triangular patch, determination of intersections between a plane where each triangular patch is located and a predetermined section, and connection of the intersections on the second image processing result, so as to obtain a third image processing result; and performing segmentation processing, classification processing, and gradient color rendering on non-triangular parts in the third image processing result in accordance with the pixel grouping condition, so as to obtain a fourth image processing result including a target rendered image.

Embodiments of the present disclosure disclose a point cloud plane segmentation method and device, electronic equipment and a storage medium, wherein the method comprises: adopting a grid division method based on phase distribution to divide a three-dimensional space corresponding to the point cloud data into a plurality of grids; performing plane fitting on points in each grid in the plurality of grids through a parallel processing manner to obtain a plurality of plane seeds; and performing neighborhood plane growth on the plurality of plane seeds through the parallel processing manner to obtain a plurality of planes corresponding to the point cloud data. The embodiments of the present disclosure can quickly and accurately determine the plurality of planes corresponding to the point cloud data, thereby improving the efficiency of point cloud plane segmentation.

The invention belongs to the technical field of point cloud models, and discloses a method and a system for identifying internal and external boundary points of a point cloud contour. The method comprises the following steps: performing plane segmentation on an original point cloud to obtain a plurality of independent plane point clouds; for each plane point cloud, obtaining a projection plane of the plane point cloud through plane fitting, extracting boundary points of the plane point cloud, and projecting the boundary points to the corresponding projection planes; converting the projected three-dimensional boundary point set to an xoy plane to obtain a two-dimensional boundary point set; constructing an envelope rectangle for the two-dimensional boundary point set by using a minimum bounding rectangle algorithm; and performing interval scanning on the two-dimensional boundary point set, calculating the distance from each two-dimensional boundary point to the corresponding rectangular edge in the scanning range, judging the two-dimensional boundary point with the minimum distance as an external boundary point, and judging the other two-dimensional boundary points as inscribed boundary points or redundant points. The method can effectively distinguish inner and outer boundary points.

Embodiments of the present application provide a point cloud segmentation method and device, electronic equipment and storage medium. The method is used to improve the accuracy of point cloud segmentation. The method comprises: obtaining point cloud data of a target region; preprocessing the point cloud data to obtain a target point cloud; wherein the number of points in the target point cloud is less than the number of points in the target region; calculating the normal vector and residual error of each point after preprocessing; wherein the normal vector is a vector represented by a straight line perpendicular to the plane where the point is located, and the residual error is used to indicate the curvature of the surface where each point is located; and performing plane segmentation on the target region according to the normal vector and residual error of each point.

This invention discloses a method and system for lane line vectorization pre-annotation based on multi-frame LiDAR point clouds. It constructs a global dense point cloud map by performing spatial coarse filtering and spatiotemporal synchronous stitching on the original point cloud data; extracts the main road surface point cloud through elevation fine filtering and plane segmentation; suppresses and enhances non-uniform backgrounds based on morphological top-hat transform; extracts candidate lane line point clouds using adaptive threshold segmentation combining statistical lower bound and recursive maximum inter-class variance method; extracts structured candidate lane lines and geometric features based on weighted iterative RANSAC line fitting guided by both intensity and longitudinal density; and constructs a global topology graph model and a multi-dimensional scoring mechanism to screen and resolve conflicts among candidate lane lines, outputting the final lane line vector map. This invention effectively overcomes uneven intensity and local noise interference, achieving high-purity and high-robustness lane line vectorization reconstruction, and significantly reducing the cost of manual annotation.

The application belongs to the technical field of three-dimensional geographic entity modeling, and proposes a building structured model topology optimization method and system based on a dynamic chain graph. Through the establishment of a geometry-topology joint optimization framework, high-precision automatic modeling in an error environment is realized, specifically including: according to input geographic scene mesh model data, building mesh data and plane segmentation results are obtained by segmentation; a dynamic chain graph structure is constructed, and a dynamic chain graph data structure is defined, wherein a node set contains double attributes of spatial coordinates and semantic weights, a chain edge set records a topological connection relationship, and a motion rule set defines four basic motion processes; a mixed energy equation is modeled, and a ternary energy equation containing a geometry fidelity term, a semantic constraint term and a topological regular term is constructed; a hierarchical optimization energy function is solved, and a three-level temperature decay optimization strategy is designed; and a building structured model is output based on the optimization result. The method can cope with the task of building structured model reconstruction in a mixed scene.

This invention relates to the field of image processing technology and discloses a planar segmentation method and apparatus for three-dimensional rendering of gradient multi-level planes. The method includes: performing grayscale conversion on a target image, and performing three-dimensional projection and pixel grouping on the first image processing result obtained from the grayscale conversion according to pixel grouping conditions to obtain a second image processing result including multiple triangular facets; performing third image processing on the second image processing result, including vertex determination for each triangular facet, intersection point determination between the plane containing each triangular facet and a preset cross section, and intersection point connection, to obtain a third image processing result; and performing segmentation processing, classification processing, and gradient color rendering on the non-triangular parts of the third image processing result according to pixel grouping conditions to obtain a fourth image processing result including the target rendered image. It is evident that implementing this invention can reduce the amount of computational data for three-dimensional rendering and improve the computational efficiency of three-dimensional rendering.

The application discloses a kind of applied to the three-dimensional model retrieval method, system, equipment, medium and product of symmetrical depth image, it is related to three-dimensional model retrieval technical field, obtain multiple standard three-dimensional models, respectively to each standard three-dimensional model is carried out symmetrical plane segmentation processing, obtain multiple target symmetrical planes, according to each target symmetrical plane and pre-trained three-dimensional feature extraction network respectively to each standard three-dimensional model is carried out three-dimensional feature extraction, obtain corresponding three-dimensional feature library, according to each target symmetrical plane and pre-trained three-dimensional feature extraction network respectively to each standard three-dimensional model is carried out three-dimensional feature extraction, obtain corresponding three-dimensional feature library. Overcome the three-dimensional model retrieval method based on two-dimensional image at present multiple dependence multi-view image or depth image carries out feature learning, there is feature extraction process complex, retrieval performance depends on network pretraining effect and so on problem, reduce the technical problem of three-dimensional model's retrieval performance and efficiency.

A computer-implemented image processing method for estimating body parameters of a particular species or breed of an animal. The image acquisition device acquires an image data set including the depth map image to obtain an image of the animal. A plane estimation algorithm analyzes the depth map image to determine a background plane behind the animal. A segmentation algorithm identifies an animal mask. Depth values outside the animal mask are reset as the distance between the viewpoint of the image acquisition device and the estimated background plane. The depth map image is normalized, limiting the data within a fixed range of numerical values, thereby generating a processed depth map image of triplex depth data (T3D) containing distance. The processed depth map image and its reset depth value for the background plane are input to a pre-trained neural network that outputs an estimate of the animal body parameter.

The application discloses a carton detection method and system based on quadrilateral fitting, and belongs to the fields of machine vision and logistics. The method comprises the following steps: simultaneously acquiring a two-dimensional color image and a three-dimensional point cloud image of a scene to be detected; inputting the two-dimensional color image into a trained anchor frame-based detection network to obtain the two-dimensional coordinates of the corner points of each quadrilateral of the carton in the two-dimensional color image; performing plane segmentation on the three-dimensional point cloud image to obtain a plurality of planes; obtaining the three-dimensional coordinates of the carton in the scene to be detected according to the two-dimensional coordinates of the corner points on the segmented planes; and the anchor frame-based detection network is trained by using a training data set {two-dimensional color image of a carton, two-dimensional coordinates of the corner points of the quadrilateral of the carton exposed in the two-dimensional color image}. The application can more accurately obtain the three-dimensional coordinates of the carton by predicting the four corner points of a certain face of the carton through the anchor frame-based detection network, and can adapt to various angles.

The application provides a disparity plane segmentation method and device based on binocular images, the method comprising: acquiring an original image and camera internal parameters, the original image comprising left and right views, a depth map and a normal vector map; generating a plane segmentation map and plane parameters according to the original image, and obtaining a true value of plane disparity according to the camera internal parameters; inputting the plane segmentation map, the plane parameters and the true value of the plane disparity into a pre-trained binocular disparity plane segmentation network to obtain a plane segmentation result output by the binocular disparity plane segmentation network, the plane segmentation result comprising a predicted plane segmentation map, a plane parameter map, a disparity map and a three-dimensional reconstruction result; wherein the binocular disparity plane segmentation network is obtained by training a pre-constructed deep neural network using image samples. The position and direction information of the plane is fully utilized, the reconstruction accuracy of the occluded area is improved, and the effect of three-dimensional reconstruction is improved.

The application relates to the technical field of electric digital data processing, and particularly discloses a steel mesh automatic recognition method based on unmanned aerial vehicle laser point cloud data, which comprises the following steps: acquiring point cloud data of a steel mesh in a tunnel or underground structure during construction, pre-processing original point cloud, plane segmentation, steel layer detection, steel axis extraction, steel direction recognition, interlayer spacing geometry parameter calculation, result output and the like. The application realizes full-automatic recognition through unmanned aerial vehicle laser scanning technology combined with an intelligent point cloud processing algorithm, improves the accuracy of a detection result, overcomes problems of low efficiency and insufficient accuracy of a traditional method which relies on manual measurement, and facilitates high-precision and high-efficiency recognition of a steel mesh in a tunnel or underground structure during construction, thereby providing a more reliable quality evaluation basis.

The invention provides a roof plane segmentation method and system fusing deep learning and local priority clustering, and the method comprises the steps: obtaining original airborne laser point cloud data of a building, and carrying out the semantic pre-segmentation of the original point cloud data based on deep learning, so as to distinguish roof surface points, plane junction points and noise points; constructing a spatial index structure to establish a local topological relation of the points, performing initial region growth on the roof surface points from the inside of the plane to the boundary by adopting a KNN local priority clustering strategy, and segmenting an initial roof plane; and for plane junction points which are not allocated in the obtained initial planes, estimating a residual threshold value of each initial plane by adopting a principal component analysis method and residual features, calculating a distance from each junction point to an adjacent plane, and completing automatic attribution allocation of the junction points according to a self-adaptive threshold value. According to the invention, building roof plane segmentation is carried out by using airborne point cloud, airborne laser radar data is used as a data source, and high-efficiency and high-precision full-automatic segmentation is realized.

The invention discloses a point cloud data-based charting method, system and device, and a storage medium. The method comprises the steps of obtaining point cloud data of a region needing charting; carrying out extraction, segmentation and boundary extraction on ground points in the point cloud data to obtain a ground contour line; extracting a building from non-ground points in the point cloud data, carrying out multi-plane segmentation on the building to extract wall point cloud, carrying out plane projection on the wall point cloud to obtain contour points of the building, and carrying out fitting to obtain a building contour line; and combining the ground contour line and the building contour line to obtain a final topographic map. Contour line recognition of buildings and structures and loading of topographic map expression conformance of the buildings and structures are achieved, and the drawing efficiency of two-dimensional line drawing is improved.

This invention provides a robot painting path planning method based on deep proximity clustering, belonging to the field of machine vision and automated robot painting technology. The method includes the following steps: First, obtain a triangular mesh model of the workpiece to be painted and perform geometric plane segmentation on the triangular mesh model; second, discretize the robot's movement space, and aggregate reachable planes into candidate continuous working ranges based on deep proximity clustering to ensure motion continuity during a single stop; then, model task selection as a set coverage problem to solve for the minimum number of stops, and model the execution order problem of tasks within the selected continuous working range as a traveling salesman problem to optimize the execution order; finally, through a two-layer global optimization of set coverage and the traveling salesman problem, generate smooth trajectory code. This invention solves the problems of low programming efficiency, long robot idle travel, and poor coating uniformity in existing large and complex workpiece painting processes, and is suitable for fully automated, high-efficiency, and high-quality painting operations on large and complex workpieces such as marine steel structures.