Method and system for processing three-dimensional special-shaped stone by robot

Abstract

The invention relates to the technical field of stone machining, in particular to a method and system for machining three-dimensional special-shaped stone through a robot, and the method comprises the following steps: obtaining path segments and extracting particle images, dividing density category identification sections, classifying the path segments, matching load conditions, screening out indexes, and mapping node coordinates to generate an interlaced section table. The method comprises the following steps: extracting a particle image, dividing density categories, identifying particle distribution differences, identifying staggered node areas in combination with spatial distribution grids, defining structural paragraphs, extracting direction vectors and sequence information to identify continuous paragraphs, adjusting attitude angle access tracks, and generating a processing path set. According to the three-dimensional special-shaped stone engraving method, combined inspection of path section direction consistency and sequence coherence is conducted, path interruption and jumping are controlled, difference analysis of the tail end posture and the engraving face direction is matched with angle adjustment, cutting offset and contour misalignment caused by posture mismatching in the machining process are reduced, and path stability and engraving coherence in the three-dimensional special-shaped stone engraving process are enhanced.

Description

Technical Field

[0001] The invention relates to the technical field of stone processing, in particular to a method and system for processing three-dimensional special-shaped stones by robots. Background Art

[0002] The field of stone processing technology includes physical processing processes such as cutting, carving, grinding, drilling, and forming for natural and artificial stones. This field covers the manufacturing process of stone products for architectural decoration, arts and crafts, garden landscaping and other purposes. The core content of this technical field is to accurately process the size, shape and texture of stone through various mechanical processing equipment or automatic control systems. The development direction of stone processing technology includes the diversification of processing technology, the intelligence of mechanical equipment and the improvement of processing accuracy. Especially in the expression of carving details and complex three-dimensional structures, CNC equipment, multi-axis robotic arms and programmable control systems have been widely used to achieve efficient processing. This field is closely related to automation technology and mechanical structure design, especially involving the complex forming requirements of irregular surfaces and large-sized stone workpieces.

[0003] Among them, the method of robot processing three-dimensional special-shaped stone refers to the method of using a multi-axis linkage industrial robot combined with digital modeling and path planning to perform three-dimensional morphological processing on natural stones with large volume and complex structure. The subject of this patent mainly focuses on technical matters such as multi-process switching, tool adaptation, and spatial path control in the stone carving process. By introducing a six-axis robot combined with a rotary worktable, continuous processes such as sawing, rough machining, semi-finishing and finishing of special-shaped stones are completed. The means used include installing different types of processing tools at the end of the robot, generating CNC path instructions based on the CAD three-dimensional model, setting processing parameters in combination with CAM software, and then the robot performs the corresponding processing actions to complete the layer-by-layer cutting and carving of the target shape of the stone. The method as a whole relies on the trajectory control of the robotic arm and the analysis of CNC path instructions to achieve three-dimensional stereoscopic forming of stone.

[0004] The existing technology mainly relies on direct analysis of the CAD model path and the issuance of CNC instructions in terms of processing path control. It lacks a real-time recognition and adaptation mechanism for the actual surface characteristics and structural density changes of the stone. In the face of uneven particle distribution on the stone surface and drastic changes in local density, it is impossible to make a dynamic response through the processing path, which often leads to excessive or too low tool load in local areas, resulting in edge collapse, overcutting and other problems. For example, when processing high-density mineral aggregation areas, if the feed depth remains unchanged, it is very easy to cause tool vibration and damage to the engraved surface. At the same time, the existing path control strategy is mostly advanced with a fixed step distance, and there is a lack of spatially distributed staggered avoidance strategies between path sections. Path overlap and posture dislocation problems are prone to occur on three-dimensional complex structures, resulting in reduced processing continuity and poor surface consistency. The end effector posture is usually only set in the initial modeling stage, and there is a lack of feedback adjustment of the real-time posture offset at the end of the path. It is impossible to ensure the directional consistency between the processing trajectory and the target engraving surface, and it is very easy to cause local relief offset or ghosting in the transition area of ​​the curved surface. The path control logic of the existing processing technology is still model-centric, ignoring the differential characteristics of the actual stone structure feedback during the processing process. Summary of the invention

[0005] In order to solve the technical problems existing in the prior art, an embodiment of the present invention provides a method for processing three-dimensional special-shaped stone by a robot, comprising the following steps: In order to achieve the above object, the present invention adopts the following technical solution: a method for processing three-dimensional special-shaped stone by a robot, comprising the following steps: S1: Obtain the robot engraving path segment, extract the particle image in the path area and analyze the arrangement state, divide the density category according to the particle spacing, and generate a particle arrangement difference segment identification set; S2: Based on the path segments marked by the particle arrangement difference segment identification set, the path segments are classified according to the particle distribution state, the tool load range is called, the path segments are mapped and matched with the load conditions, the path index is extracted, and the path segments without the feed depth set are classified into the adjustment range to obtain the adjustable feed operation path segment sequence; S3: Based on the path segments in the adjustable feed operation path segment sequence, the path segment node coordinates are mapped to the spatial distribution grid, the number of grid area nodes is counted, and a staggered operation path segment registration table is obtained; S4: based on the path line group in the staggered operation path segment registration table, extract the path point direction vector and verify the direction continuity, analyze the path continuity in combination with the connection sequence, and obtain the continuous processing segment group; S5: Based on the path direction of the segments in the continuous processing segment group, the difference between the posture of the end effector of each path segment and the direction of the engraving surface is analyzed, and the angle of the deviation segment is adjusted and then connected to the trajectory to obtain a special-shaped stone processing path set.

[0006] As a further solution of the present invention, the particle arrangement difference segment identification set includes particle arrangement direction data, average particle spacing interval, dense and sparse area boundary marks, the adjustable feed operation path segment sequence includes a path segment number set, a load adaptation annotation item, and a feed depth setting identifier, the staggered operation path segment registration table includes the intersection node space coordinates, node density distribution indicators, and staggered path segment index numbers, the continuous processing segment group includes direction continuous path segments, node sequence mapping relationships, and stable continuation segment numbers, and the special-shaped stone processing path set includes posture adjustment path segments, processing trajectory sequence indexes, and end effector angle correction values.

[0007] As a further solution of the present invention, the specific steps of S1 are: S101: Obtain the robot path segment in the stone carving task, extract the stone surface image in the path corresponding area, identify the particle arrangement direction and particle spacing distribution, extract the particle position and arrangement characteristics in the path segment, and obtain the particle arrangement recognition result; S102: Based on the particle arrangement recognition result, analyzing the connectivity of the particle directions in the path segment, combining the particle spacing variation trend and the quantity distribution range, classifying the path segment according to a preset classification standard, and generating a particle density segment classification result; S103: calling the particle density segment classification result, marking the path points of the difference boundaries in the path segment, updating the density distribution state in the path index and structure information, and obtaining a particle arrangement difference segment identification set.

[0008] As a further solution of the present invention, the specific steps of S2 are: S201: Based on the path segments marked in the particle arrangement difference segment identification set, the path segments are divided according to the continuity of the particle distribution state and the density distribution trend in the path segments, and the path segments are classified into differential category identifications according to the particle number distribution value and the spacing mean, so as to obtain a path segment classification label set; S202: Based on the path segment classification mark set, read the load range value set on the tool structure, extract the total length of each type of path segment and correspond it to the category number, calculate the total execution length ratio of the category path segment, filter the matching path segment numbers, and obtain the executable path segment index table; S203: calling the executable path segment index table, detecting the record items in the path number that do not have the feed depth set, screening the number and merging and marking it with the original index, re-arranging the data index according to the merged path set, and obtaining an adjustable feed operation path segment sequence.

[0009] As a further solution of the present invention, the calculation formula of the total execution length ratio of the category path segment is specifically: ; in, Represents the total execution length ratio of the path segment of category i, represents the total execution length of the i-th path segment, represents the arithmetic mean of the load values ​​set for all path segments in the i-th path segment, represents the median of the load values ​​set for all path segments in the i-th path segment. represents the load weight coefficient of the structure corresponding to the i-th path segment, Represents the arithmetic mean of the total execution length of all class path segments, Represents the total execution length of the k-th path segment, Represents the average value of the load values ​​set for all path segments in the k-th path segment. represents the structural load weight of the k-th path segment, and K represents the total number of path segment categories.

[0010] As a further solution of the present invention, the specific steps of S3 are: S301: based on the path segments in the adjustable feed operation path segment sequence, extracting node coordinate data in the path segment, dividing the node position into a spatial distribution grid according to the coordinate range, identifying the grid number to which the node belongs and establishing a mapping index, and obtaining a node grid belonging relationship table; S302: Based on the node grid attribution relationship table, the total number of nodes associated with the grid number and the corresponding coordinate boundary are extracted, the node spatial distribution density value within the unit area under the number is calculated, and the grid number whose distribution density exceeds the reference value is selected in combination with the density reference standard to generate a high-density node grid number set; S303: calling the high-density node grid number set, filtering the path segment index information associated with the number set, merging the corresponding path segments into group data according to the grid numbers, and adding interleaved area marks to obtain an interleaved operation path segment registration table.

[0011] As a further solution of the present invention, the calculation formula of the node spatial distribution density value per unit area under the number is specifically: ; in, Represents the node spatial distribution density value per unit area under the grid number z. represents the total number of nodes associated with the grid number z, represents the qth weighting factor associated with the node distribution, Represents the relative spatial difference between the grid number z and the qth node, Represents a coordinate value of the sth node in the grid number z, Represents the average value of the coordinates of all nodes in the grid number z. represents the boundary area corresponding to the grid number z, Representative The coefficient of the boundary influence factor of the term, Represents the grid number z. The boundary morphology adjustment parameters are u, the total number of factors involved in the weighted calculation of node space difference, v is the total number of nodes contained in the grid number z, and w is the number of types of factors affecting the grid boundary.

[0012] As a further solution of the present invention, the specific steps of S4 are: S401: based on the path line groups in the staggered operation path segment registration table, extract the direction vector information of adjacent points in each path, sort the vector arrangement order according to the path node order, analyze the consistency of the direction change trend between adjacent vectors, and obtain the path direction continuous segment group; S402: Based on the path direction continuous segment group, extract the arrangement order of the corresponding path points in the paragraph, detect whether there is a sequence interruption or reverse jump of the path points, and select the path segment information that maintains a continuous point sequence to obtain a path point sequence continuous segment group; S403: calling the path point sequential continuous segment group, performing matching processing against the existing directional continuity information, verifying the path line group that satisfies both directional stability and point arrangement continuity, and obtaining a continuous processing segment group.

[0013] As a further solution of the present invention, the specific steps of S5 are: S501: based on the path direction of the path segments in the continuous processing segment group, extract the end effector posture data, and compare it with the corresponding engraving plane direction, identify the path segments whose spatial direction difference reaches the posture deviation reference value, and obtain the posture deviation meeting the condition segment index set; S502: calling the index set of the segment whose posture deviation meets the condition, reading the end angle information of the corresponding path segment, adjusting the posture angle according to the preset rule, and binding the adjustment value to the path segment posture information to generate a posture correction path segment set; S503: Based on the posture correction path segment set, the correction segment is added to the processing trajectory data, the path segment number and the sequence index are updated, and the processing trajectory is adjusted to obtain the special-shaped stone processing path set.

[0014] The robot system for processing three-dimensional special-shaped stones includes: The path construction module obtains the robot processing path segment of three-dimensional special-shaped stone carving, collects the particle image of the stone surface in the area covered by the path segment, partitions the particle arrangement direction, spacing distribution and boundary changes in the image, delineates the structural distribution inside the path segment according to the particle concentration degree, and generates a particle arrangement difference area identification set; The structure recognition module identifies the particle quantity distribution and arrangement spacing of the concentrated path segment based on the particle arrangement difference area, identifies the particle dense, moderate and sparse segments, reads the load-bearing interval marked in the tool structure equipped with the robot, matches the particle category and load condition to which the path segment belongs, extracts the path segment number that meets the processing requirements, checks the feed depth setting status for each numbered path, screens out unset items and classifies them into the adjustment path number, merges the numbers and reorders them, and obtains an adjustable feed operation path segment sequence; The load matching module locates the intersection coordinates of the nodes in the path segment based on the path segments included in the adjustable feed operation path segment sequence, delimits the spatial area according to the node coordinates and maps them to the spatial distribution grid, extracts the node quantity data in the area, compares the density benchmark value to select the area number whose node density reaches the benchmark value, and associates the corresponding path segment number to obtain the staggered operation path segment registration table; The segment analysis module extracts the direction vector sequence of continuous path points in the segment based on the path segments in the staggered operation path segment registration table, identifies the line segments with continuous direction changes, analyzes the consistency of the point arrangement sequence, and selects the path segment numbers with both direction consistency and point continuity to obtain a continuous processing segment group; The posture adjustment module calls the end effector posture value of the path segment in the continuous processing segment group, extracts the carving surface direction of the path segment, compares the posture angle relationship between the two, screens out the path segment whose offset exceeds the reference angle, corrects the posture angle and inserts the corrected segment into the original processing trajectory to obtain the special-shaped stone processing path set.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are: In the present invention, by extracting particle images and dividing density categories, identifying particle distribution differences, combining spatial distribution grids to identify staggered node areas, defining structural segments, jointly checking path segment direction consistency and sequence continuity, controlling path interruptions and jumps, and analyzing the difference between the end posture and the carving surface direction and coordinating the angle adjustment, the cutting offset and contour misalignment caused by posture mismatch during the processing process are reduced, and the path stability and carving continuity during the three-dimensional special-shaped stone carving process are enhanced. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0017] Figure 1 It is a schematic diagram of the steps of the present invention; Figure 2 It is a system module diagram of the present invention. DETAILED DESCRIPTION

[0018] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0019] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "example" in the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of the word "example" is intended to present the concept in a specific way. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or it can be either of the two.

[0020] In the embodiments of the present invention, "image" and "picture" can sometimes be used interchangeably. It should be noted that when the difference between them is not emphasized, the meanings they intend to express are the same. "of", "corresponding, relevant" and "corresponding" can sometimes be used interchangeably. It should be noted that when the difference between them is not emphasized, the meanings they intend to express are the same.

[0021] In the embodiments of the present invention, sometimes the subscripts such as W 1 It may be written in non-subscript form such as W1. When the difference is not emphasized, the meaning is the same.

[0022] In order to make the technical problems, technical solutions and advantages to be solved by the present invention more clear, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0023] See also Figure 1 The method for processing three-dimensional special-shaped stone by robot comprises the following steps: S1: Obtain the path line segment preset by the robot in the stone carving task, extract the surface particle image of the stone corresponding to the path in the robot running section, analyze the particle arrangement direction, average particle spacing, and boundary changes between dense and sparse areas, divide the path segment into density categories according to the average particle spacing and the number of particles, group and mark the partition boundary points, and update the path data index to generate a particle arrangement difference segment identification set; S2: Based on the path segments marked in the particle arrangement difference segment identification set, the path segments are classified according to the particle distribution state to which the path segments belong, the load range preset on the tool structure is read, the path segment classification results are mapped and matched with the corresponding load conditions, and the path segment indexes that meet the execution requirements are extracted according to the matching results. The path segments that have not completed the feed depth setting are synchronously included in the adjustment range to obtain the adjustable feed operation path segment sequence; S3: Based on the path segments in the adjustable feed operation path segment sequence, the intersection node positions in the path segment are divided into spatial distribution grids according to coordinates, the number of path nodes in each grid unit is counted, the grid number sequence is called to compare with the node density data, and the units with a number exceeding the density threshold are marked as staggered segment areas, and the corresponding path segment numbers are summarized to obtain the staggered operation path segment registration table; S4: Based on the path line groups in the staggered operation path segment registration table, the direction vectors of adjacent points in each path are sequentially extracted, and the sections with consistent directions are identified based on the direction continuity between the path points. The length of the line segments between the connection points is compared with the sequence of the path points to identify the set of paths with stable directions and continuous points, and obtain a continuous processing section group; S5: Based on the path direction of the segments in the continuous processing segment group, the posture deviation angle between the current posture of each end effector and the engraving plane direction is analyzed, and the posture angle of the path segment whose deviation angle reaches the set range is adjusted. The adjusted path segment is added to the engraving trajectory to obtain the special-shaped stone processing path set.

[0024] The particle arrangement difference segment identification set includes particle arrangement direction data, average particle spacing interval, dense and sparse area boundary markers, the adjustable feed operation path segment sequence includes path segment number set, load adaptation annotation items, and feed depth setting identification, the staggered operation path segment registration table includes the intersection node space coordinates, node density distribution indicators, and staggered path segment index numbers, the continuous processing segment group includes direction continuous path segments, node sequence mapping relationships, and stable continuation segment numbers, and the special-shaped stone processing path set includes posture adjustment path segments, processing trajectory sequence indexes, and end effector angle correction values.

[0025] The specific steps of S1 are: S101: Obtain the robot path segment in the stone carving task, extract the stone surface image in the path corresponding area, identify the particle arrangement direction and particle spacing distribution, extract the particle position and arrangement characteristics in the path segment, and obtain the particle arrangement recognition result; The path segment data group recorded in the engraving task work order is called, and the spatial coordinate starting point and end point of each segment are extracted one by one, and its spatial positioning range on the robot processing workbench is calculated, and then the positioning link of the robot operation section is entered. The stone surface image is collected by the visual acquisition device in this spatial area, and the image resolution is set to 12 pixels per millimeter to ensure that the particles can be clearly identified. In the collected image, the channel brightness value distribution is first used for preliminary segmentation, and the particle dot matrix with grayscale values ​​in the range of 65 to 180 in the image is extracted. Then, the contour tracking operation is performed on the edge of the particle to determine the complete boundary area of ​​the particle, and the minimum circumscribed rectangle of the contour is obtained to measure the actual spacing between the particles. On this basis, the center point coordinates of all particles are identified, and then the path segment is projected into the acquisition image coordinate system, and the particle center point data within the coverage range of the path segment is selected as the particle set in the current path segment. On this basis, the direction of the connection between the particle center points is calculated, and the pairwise connection direction vectors are defined, and their angles are statistically divided. The direction with an angle change of less than 15 degrees is selected as the main arrangement direction. For example, among the 80 connecting line segments formed by the extracted particle points in a certain path segment, 65 groups of angles are distributed between 5 and 15 degrees, and the main arrangement direction is determined to be the 10-degree direction. Subsequently, the center point coordinates of all particles are parallel projected according to their arrangement direction, and the distance between adjacent projection points is measured as the particle spacing value. The statistical results show that the spacing is concentrated between 2.3mm and 3.1mm, which is recorded as the average particle spacing in the path segment as 2.7mm. Subsequently, all the particle points on the path segment are extracted, and the number of particles per unit path length is calculated. Combined with the average particle spacing, the path segment is classified as a high-density arrangement segment, and the start and end coordinates corresponding to the path segment are written into the structure as partition boundary points, marking the position index number of the path segment in the path set, and returning the particle arrangement direction, average particle spacing and particle number as record items, updating the structure attribute field of the path segment, and obtaining the particle arrangement recognition result.

[0026] S102: Based on the particle arrangement recognition result, the connection of the particle directions in the path segment is analyzed, and the path segment is classified according to the preset classification standard in combination with the particle spacing change trend and the quantity distribution range, so as to generate the particle density segment classification result; First, the sequence of main arrangement direction values ​​identified in each path segment is extracted, and the range of particle arrangement angle variation is divided according to the arrangement continuity of particles within the path segment. In any path segment, if there are three adjacent particles with arrangement direction angles of 9°, 11°, and 13°, respectively, they are considered to have stable directional connectivity. If the difference between adjacent angles is greater than 10°, such as 17°, 28°, and 39°, they are recorded as areas with poor directional connectivity, and areas with good directional connectivity are marked as stable segments. Secondly, the particle spacing values ​​of each path segment are arranged in sequence to construct a variation trend sequence, and the difference between adjacent spacings is obtained. If the spacing difference between multiple consecutive particle pairs in the path segment is less than 0.4 mm, it is considered to be a segment with a stable spacing trend. If a segment shows a regular increase or decrease in the variation trend, it is considered to show a trend fluctuation in the spacing. The particle spacing trend type is recorded in units of path segments, and then The distribution of the number of particles is judged, and the number of particles per unit length (10mm interval) on each path segment is calculated. If the number of particles per unit length is between 4 and 6, it is defined as a medium-density segment, if the number exceeds 6, it is defined as a high-density segment, and if the number is less than 4, it is defined as a low-density segment. The judgment standard comes from the statistics of the average particle distribution interval of 200 stone samples. In actual implementation, if a path segment is 30mm long and contains 18 particles, the average number of particles per unit length is 6, which is classified as a high-density segment. Subsequently, the arrangement direction stability, spacing change trend and unit length particle density on the path segment are integrated. The path segment with two of the three conditions that meet the stability standard is classified as a consistent density segment. If two of the three items fluctuate, it is classified as a density fluctuation segment. The division results of various path segments are accompanied by the original path segment number and boundary coordinates to generate a particle density segment classification result.

[0027] S103: calling the particle density segment classification result, marking the path points of the difference boundary in the path segment, updating the density distribution state in the path index and structure information, and obtaining the particle arrangement difference segment identification set; Call the calibrated path segment category label and boundary information in the particle density segment classification result, read the index number of the starting point and the end point of each path segment in the processing path, extract the path points at the junction of different density categories, extract the particle density value labels of adjacent points within the path segment according to the point sequence index, judge whether the label change is continuous and non-repetitive, and record the spatial coordinates of the point at the jump from medium density to high density or from low density to medium density. In the stone carving task example, if the continuous point density label sequence in the path segment numbered P32 is 2, 2, 3, 3, then extract the coordinates of the middle point at the junction between the second 2 and the first 3 as the identification basis of the difference boundary path point. After completing the extraction of the boundary path point, press The original path index list is updated with the original path number and point sequence, and the path segment length, point coordinates, direction vector and other structural items in the original path structure information table are merged with the newly added density distribution state. The fusion method uses the path segment number as the associated key value, and the density interval number obtained in the particle density segment classification is written into the structure information table according to the path segment number, so that each path segment structure information is accompanied by its corresponding density segment type number and difference boundary point coordinate information. For example, the density category of the path segment numbered P45 is category 3, and the boundary point position is between the 27th and 28th points. The newly added field record in the path structure table is "density category 3, boundary points 27-28", and finally the particle arrangement difference segment identification set is obtained.

[0028] The specific steps of S2 are: S201: based on the path segments annotated in the particle arrangement difference segment identification set, the path segments are divided according to the continuity of the particle distribution state and the density distribution trend in the path segments, and the path segments are classified into differential category identifications according to the particle number distribution value and the spacing mean, so as to obtain a path segment classification label set; First, the number of each path segment, the corresponding particle distribution data and its arrangement state in the spatial coordinates are read, the coordinate set of all particles in the path segment is extracted, and the particle positions are sorted in the direction of the path segment length to construct a particle arrangement sequence. The spacing change sequence between adjacent particles in the path segment is calculated, and the number of spacing fluctuations in different particle segments is counted. Continuity is evaluated based on the proportion of fluctuations and the span of the variation interval. If the spacing difference in a continuous segment exceeds 40% of the maximum spacing of the total segment, it is recorded as a discontinuous state, otherwise it is defined as a continuous distribution state. For example, the path number P12, the adjacent spacings of particles in this segment are 1.2, 1.3, 1.4, 2.7, 1.3, and 1.2, respectively. The value 2.7 in the middle exceeds twice the average value of 1.28, which is a typical segment with continuity interruption. Combined with the continuity classification The analysis results further call the total number of particles and segment length parameters of each path segment, and calculate the particle density value for each segment. The density value is defined as "the number of particles per unit path length". For example, if the path segment length is 30 mm and the number of particles is 15, the density is 0.5 particles / mm. Combined with the variation of the particle density distribution of each path segment, the low, medium and high density division intervals are set by comparison. For example, low density is set to less than 0.3 particles / mm, medium density is 0.3 to 0.7 particles / mm, and high density is greater than 0.7 particles / mm. A mapping is established between the density value in the path segment and the density partition. Finally, the corresponding density classification label and continuity mark are added to the path segment information structure. For example, the path segment numbered P12 is marked as "density category: medium density, continuity state: discontinuous", thereby obtaining the path segment classification mark set.

[0029] S202: Based on the path segment classification mark set, read the load range value set on the tool structure, extract the total length of each type of path segment and correspond it to the category number, calculate the total execution length ratio of the category path segment, filter the matching path segment numbers, and obtain the executable path segment index table; The calculation formula for the total execution length ratio of the category path segment is as follows: ; in, Represents the total execution length ratio of the path segment of category i, represents the total execution length of the i-th path segment, represents the arithmetic mean of the load values ​​set for all path segments in the i-th path segment, represents the median of the load values ​​set for all path segments in the i-th path segment. represents the load weight coefficient of the structure corresponding to the i-th path segment, Represents the arithmetic mean of the total execution length of all class path segments, Represents the total execution length of the k-th path segment, Represents the average value of the load values ​​set for all path segments in the k-th path segment. represents the structural load weight of the k-th path segment, where K represents the total number of path segment categories; Assume the following values: =300 m, =1.5 times, =1.4 times, =0.8, =300 meters, =300 meters, =1.5 times, =0.8, K=5; Example of calculation derivation process: calculate ; ; ; calculate ; ; =1599; Calculate the final result : ; Result interpretation: The result shows that the total execution length of the i-th path segment accounts for 5.31%, which means that its execution resource share in all categories of path segments is only 5.31% of the total. This value is determined by the path segment length, structural load weight, and load change difference. The denominator reflects the weighted execution total of all categories of path segments. This result is used to measure the execution importance of the i-th path segment in the overall path system and is a key basis for path screening and priority scheduling. When the proportion is low, this type of path segment will be given a lower execution priority in task allocation.

[0030] S203: calling the executable path segment index table, detecting the record items in the path number that do not have the feed depth set, screening the number and merging and marking the original index, re-arranging the data index according to the merged path set, and obtaining the adjustable feed operation path segment sequence; First, all the path segment numbers are read one by one, and the feed parameter field corresponding to each number item is checked to confirm whether there is a set feed depth value field content. If the field value is empty or zero, it is determined that the numbered path segment has not set an effective feed depth. The number is recorded and included in the unset path number set. For example, if the feed field content of numbers P118, P127, and P134 is empty during the reading process, they are included in the unset number set. Subsequently, the number set is cascaded with the original executable path segment index table, and the path number is used as the primary key. The original field content and the newly marked field "Depth Status: Unset" are spliced ​​and attached to form data with a complete structural description. Then, based on the path number sequence, all data rows are rearranged in ascending order of number. In the process, the original order field is retained and the new index sequence field is updated to form a unified numbering system. On this basis, all numbered path segments marked as "unset" and their segment positions are extracted and added as the data source of the feed parameter adjustment list. For example, after the rearrangement, the corresponding indexes of numbers P102, P105, and P108 are 12, 17, and 23 respectively, then the new record in the rearrangement structure is "path P102, index 12, status unset". Finally, a processing path segment structure list is constructed with the path number as the primary key and the index sequence as the sorting basis, and an adjustable feed operation path segment sequence is obtained.

[0031] The specific steps of S3 are: S301: based on the path segments in the adjustable feed operation path segment sequence, extract the node coordinate data in the path segment, divide the node position into a spatial distribution grid according to the coordinate range, identify the grid number to which the node belongs and establish a mapping index, and obtain a node grid belonging relationship table; First, extract all the node coordinate information in each path segment, traverse the spatial position of the node, and parse the coordinate data into the X, Y, and Z axis coordinate value sets in the three-dimensional space. Then, set the spatial grid boundary according to the working range of the stone processing platform, and define the minimum segmentation unit within the three-dimensional range. Each unit volume is 10mm×10mm×10mm, and is assigned a unique number. Perform position matching operation on each node coordinate, compare the X, Y, and Z coordinate values ​​of each node with the grid boundary respectively, and determine the spatial grid number interval in which the coordinate point falls. After matching, form a one-to-one mapping relationship between the node number and the corresponding grid number. For example, there is a node in the path segment numbered P012, and its coordinates are (135.2 , 87.6, 42.3), falls into the grid unit numbered (13, 8, 4) under the condition of 10mm unit grid, then the node number and the grid number are mapped as "P012-N03→G13-08-04". In this way, the spatial grid ownership of all nodes in all path segments is recorded item by item, and a list structure is formed. The mapping result structure is further grouped and archived according to the path segment number, and all node grid numbers in each group of path segments are included in a group of data structures. The index sequence of the path segment and the node position in the mapping structure is recorded to ensure that the spatial position identifier corresponding to each node can be quickly traced according to the path segment in subsequent operations. After completing the above operations, a node grid ownership relationship table is obtained.

[0032] S302: Based on the node-grid attribution relationship table, the total number of nodes associated with the grid number and the corresponding coordinate boundary are extracted, the node spatial distribution density value within the unit area under the number is calculated, and the grid number with a distribution density exceeding the reference value is screened in combination with the density reference standard to generate a high-density node grid number set; The calculation formula for the node spatial distribution density value per unit area under the number is as follows: ; in, Represents the node spatial distribution density value per unit area under the grid number z. Represents the total number of nodes associated with the grid number z, represents the qth weighting factor associated with the node distribution, Represents the relative spatial difference between the grid number z and the qth node, Represents a coordinate value of the sth node in the grid number z, Represents the average value of the coordinates of all nodes in the grid number z. represents the boundary area corresponding to the grid number z, Representative The coefficient of the boundary influence factor of the term, Represents the grid number z. The boundary shape adjustment parameters, u is the total number of factors involved in the weighted calculation of node space difference, v is the total number of nodes contained in the grid number z, and w is the number of types of factors affecting the grid boundary; This formula is used to calculate the spatial distribution density of nodes in a specific area. The following is a detailed explanation of each part of the formula and the calculation derivation: Assume the following values: = node, =0.8, =1.2, =2.5 m, =3.0 m, =50.0 m, =52.5 m, =51.25 m, =100 square meters, =1.5, =1.0, =0.3 =0.2; Formula calculation example: Consider the case of u=2, v=2, w=2, the sum of weighted node space differences: ; Mean square error of node coordinate values: ; The square root is: ; Boundary Adjustment Sum: ; Substituting these values ​​into the density formula: ; This result indicates that the calculated The spatial distribution density value of the nodes per unit area in the grid number z is 1.225, indicating that the nodes are relatively dense in this area, reflecting the dense distribution of nodes in the grid area, based on which we can further decide on the optimization or enhancement measures of network resources.

[0033] S303: calling a high-density node grid number set, filtering out path segment index information associated with the number set, merging corresponding path segments into group data according to the grid number, and adding an interleaved area mark to obtain an interleaved operation path segment registration table; First, the matching operation is carried out on each numbered data in the order of spatial number, and all node items corresponding to the grid number in the node grid attribution relationship table are obtained. The path segment number information bound to these node items is extracted, and the path segments corresponding to multiple nodes with the same grid number are uniformly classified to build an aggregation relationship structure between the path segment and the grid number. In the aggregation process, the path segment index is set as the first-level identifier and the grid number is set as the second-level grouping key value. The multiple grid numbers associated with each path segment are summarized, and a new field is added to the path segment structure to record the grid number set to which it belongs. For example, the nodes contained in the path segment numbered P204 belong to the numbers G12-04, G12-05, and G13-04, and the grid distribution recorded in its path segment structure is [G12-04, G12-05, and G13-04]. -05, G13-04], then determine whether it belongs to the high-density grid area, and filter according to whether the number appears in the high-density node grid number set. If any number appears in the number set, the path segment is marked as an interlaced area segment, and the interlaced area mark field "Interlaced identification: Yes" is added, otherwise the mark "Interlaced identification: No" is retained. After the path segment marking is completed, all path segment index information that has been classified into the interlaced area is uniformly output and arranged in ascending order according to the path segment number. Each record is accompanied by the path segment number, number of nodes, grid number distribution and interlaced identification field value. For example, the number P204 corresponds to the output of "P204 | 6 nodes | G12-04, G12-05, G13-04 | Interlaced identification: Yes". Finally, all marking results are summarized to generate an interlaced operation path segment registration table.

[0034] The specific steps of S4 are: S401: based on the path line groups in the staggered operation path segment registration table, extract the direction vector information of adjacent points in each path, sort the vector arrangement order according to the path node order, analyze the consistency of the direction change trend between adjacent vectors, and obtain the path direction continuous segment group; First, the continuous node points in each path segment are traversed, the coordinate values ​​of each pair of adjacent nodes are extracted and arranged in the order of the path, and the direction vector is constructed for any two adjacent nodes. The direction vector is defined as the three-dimensional coordinate difference of the latter node minus the former node, and a complete vector sequence inside the path segment is sequentially constructed. Then, the direction change trend of the vector sequence is analyzed, and each group of continuous vector pairs is used as the analysis unit. The angle value is calculated and compared with the continuity judgment threshold. If the angle values ​​of three consecutive groups of vectors are all within 15°, they are recorded as a sequence segment with consistent direction. On the contrary, if any of the angles is greater than 30°, it is considered as a direction interruption. In the path numbered P212, there are nodes N01 to N06, and the angle values ​​of their adjacent vectors are 12°, 14°, 13°, 3°, respectively. 5°, 17°, it can be identified that the first three segments constitute a directional continuous segment, and the fourth segment has a directional mutation point, and the directional continuity is interrupted. All node sequences with three or more consecutive consistent directions are identified in the path segment, and their start and end node numbers and index positions in the path segment are extracted. At the same time, the segments are grouped according to the path segment number. Each group of data contains the start node, the end node, the segment direction vector sequence and the original path number identifier. The directional continuous segment is combined with the interlaced area mark to which the path segment belongs to establish a data structure. For example, N01 to N04 in number P212 are identified as continuous segments, which are recorded as "P212|N01-N04|angle sequence【12,14,13】". Finally, all segments that meet the directional continuity requirements are output to obtain a path direction continuous segment group.

[0035] S402: based on the path direction continuous segment group, extract the arrangement order of the corresponding path points in the paragraph, detect whether there is a sequence interruption or reverse jump of the path points, select the path segment information that maintains a continuous point sequence, and obtain the path point sequence continuous segment group; First, read the numbering sequence information of the path points contained in each segment one by one, extract the index values ​​of the starting node and the ending node in the original path segment, and construct a node sequence by arranging the indexes from small to large. Then, the continuity of the adjacent node index values ​​in the node numbering sequence is judged. If the index difference between two adjacent node numbers is 1, the sequence is considered to be continuous, otherwise it is marked as a sequence interruption. For example, the node sequence in the segment numbered P308 is N12, N13, N15, and N16, then a jump between N13 and N15 is identified, and the interruption position is recorded. If there is a reverse order, such as the node numbering sequence is N25, N24, and N23, the continuity is During the judgment, it is identified that the index value shows a decreasing trend and is classified as a reverse jump type. The integrity of the node number sequence in each path segment is used for screening, and all fragments with increasing indexes and no jumps are retained as valid sequence fragments. The fragments are grouped and sorted according to the path number, and the path number, the start and end node numbers of the fragment, and the continuity judgment status are recorded in the data structure. For the example numbered P308, only N15 to N16 constitute a continuous fragment, which is finally included and output in the format of "P308 | N15-N16 | Sequence status: continuous". All path segment fragments that meet the sequence integrity conditions are summarized to obtain a path point sequence continuous fragment group.

[0036] S403: calling the path point sequential continuous segment group, performing matching processing against the existing direction continuity information, verifying the path line group that satisfies both the direction stability and the point arrangement continuity, and obtaining the continuous processing segment group; First, it is necessary to number and identify each path point in the path line, extract the spatial coordinate value (x, y, z) and topological order information of each path point in the graphic database or three-dimensional geometric model, and form a sequential set of path points. By analyzing the set point by point, the connection status between adjacent path points is identified, and the distance, direction vector and angle information between each two path points are recorded. On this basis, continuous segments are extracted in sequence. Each segment must meet the requirements of continuous path point numbering, spatial distance less than the set segment length threshold δ, and direction angle less than the set direction change threshold θ. For example, there are path points P1 to P100 on a curved path, the segment length threshold δ is set to 3.5mm, and the direction change threshold θ is 12°. If the distance between adjacent path points between P1 and P10 is less than 3.5mm, and the direction angle gradually changes by no more than 12°, then P1 to P10 can be called as a sequential continuous segment group, and then matched against the existing direction continuity information. It is necessary to read the historical data. The directional vector template information in the historical processing paragraph database is used to perform dot product operation on each directional vector in the calling fragment group, and the similarity s_i=cos(θ_i) is calculated, where θ_i is the angle between the current paragraph directional vector and the template directional vector. If the similarity s_i>0.98, the direction is determined to be continuous. In practice, if the template vector is (1, 0, 0) and the current paragraph direction is (0.998, 0.05, 0), the calculated angle θ_i is about 2.86°, s_i≈0.9987, which meets the directional continuity requirements. To further verify the path line group of directional stability and point arrangement continuity, all continuous paragraphs need to be combined into path line groups in sequence, and the directional change amplitude Δθ between adjacent paragraphs is compared in turn. If Δθ<θ (set the directional stability threshold θ, for example, 10°), and the distance between the starting and ending points of the paragraph is less than δ (set the paragraph continuity threshold, for example, 5mm), it is determined to meet the continuity, and finally a continuously processed paragraph group is obtained.

[0037] The specific steps of S5 are: S501: based on the path direction of the path segments in the continuous processing segment group, extract the end effector posture data, and compare it with the corresponding engraving plane direction, identify the path segment whose spatial direction difference reaches the posture deviation reference value, and obtain the posture deviation meeting the condition segment index set; First, extract the current 3D posture data of the end effector of each path segment, obtain the posture angle information of the end effector in the X, Y, and Z directions, and combine it with the reference engraving plane direction vector set in the robot body coordinate system. Match the engraving plane normal direction value corresponding to each posture point from the starting point to the end point of the path segment in sequence. In the matching process, establish a spatial difference judgment method between the posture angle vector and the engraving surface direction vector, and judge the spatial angle between each posture vector and the direction vector of the engraving surface of the node to which it belongs. If the angle value of the two vector directions under the unit vector projection exceeds the set posture deviation reference value interval, the current path segment number and the corresponding node index number can be recorded. If there are nodes N01 to N08 in the path segment numbered P218, and the angle offset between nodes N03 and N04 in the Z-axis direction exceeds the attitude deviation reference value of ±10 degrees, then P218 is marked as a segment with attitude deviation, and N03-N04 is recorded as the critical node interval of deviation. After performing this type of judgment operation on all path segments, the path segments that do not meet the set deviation value requirements are eliminated, and only the path segment information that meets the attitude deviation judgment is retained, and its path number, start and end node number, deviation type, offset direction and other information are included in the structured list and output, and all path segments that meet the attitude deviation reference requirements are summarized to obtain an index set of conditional segments with attitude deviation.

[0038] S502: calling the index set of the segments that meet the posture deviation conditions, reading the end angle information of the corresponding path segment, adjusting the posture angle according to the preset rules, and binding the adjustment value to the path segment posture information to generate a posture correction path segment set; Call the path segment number and the corresponding node position data, extract the angle posture information of the end effector of each path point in the segment from the path data structure, obtain the Euler angle or posture angle record of the corresponding point according to the XYZ three-axis direction, extract the current posture angle at point N03 as 8 degrees for the X axis, 2 degrees for the Y axis, and 15 degrees for the Z axis, look up the set posture adjustment rule table, obtain the Z axis direction offset allowable range of ±5 degrees, compare the Z axis offset of 15 degrees and the offset reference value difference of 10 degrees, which is higher than the set tolerance interval, record the current node as the node to be adjusted, call the matching adjustment rule, correct the Z axis direction offset angle, and set the adjusted angle to 10 degrees , the Z-axis correction value is recorded as 5 degrees, and the Z-axis angle correction field is inserted into the node record structure. After updating, the corrected posture of node N03 is X8, Y2, Z10, and the subsequent points such as node N04 repeat the operation process in sequence. According to the rule table, the correction angles of all posture deviation nodes are added and recorded. After completing the full node correction of a single path segment, the posture field information in the path segment data structure is updated, and the original posture and the corrected posture are written into the posture change field as comparison records, and the path segment number is recorded as the posture corrected segment. Finally, all the data path segments that have completed the angle correction are sorted by the original numbers and output as a set to generate a posture correction path segment set.

[0039] S503: Based on the posture correction path segment set, the correction segment is added to the processing trajectory data, the path segment number and the sequence index are updated, and the processing trajectory is adjusted to obtain the special-shaped stone processing path set; First, the path segment structure content is read from the original processing trajectory data set, the corrected path with segment number P042 is identified, the insertion position corresponding to the original P042 is located in the processing trajectory structure, the corrected segment is added to the trajectory data structure by sequential insertion, and the original segment data is replaced by it, the original segment start and end point numbers are retained and the number is synchronously updated to P042-A, the path index field is reassigned, the subsequent path segment numbers are sequentially updated according to the numbering sequence P043 to P048, and the segment number mapping table content recorded in the trajectory structure is synchronously adjusted, the old number in the index mapping table after the corrected segment is inserted is deleted, and the new number is matched with the corresponding segment data, and the segment number is written into the trajectory index total table after confirming that there is no conflict in the segment number, and the path segment sequence is traversed and checked. After confirming that the logic of the number increment of each segment is consistent with the node sequence, a trajectory segment sequence confirmation record is generated, the numbers and node sets of all inserted segments are extracted, and a trajectory segment update table is constructed, and a complete trajectory set data containing corrected posture segments, number sequence updates, and trajectory structure adjustments is output to obtain a special-shaped stone processing path set.

[0040] See also Figure 2 , a robot system for processing three-dimensional special-shaped stones, including: The path construction module obtains the robot processing path segment of three-dimensional special-shaped stone carving, collects the particle image of the stone surface in the area covered by the path segment, partitions the particle arrangement direction, spacing distribution and boundary changes in the image, delineates the structural distribution inside the path segment according to the particle concentration degree, and generates a particle arrangement difference area identification set; The structure recognition module identifies the particle quantity distribution and arrangement spacing of the concentrated path segment based on the particle arrangement difference area, identifies the particle dense, moderate and sparse segments, reads the load-bearing interval marked in the tool structure equipped with the robot, matches the particle category and load condition to which the path segment belongs, extracts the path segment number that meets the processing requirements, checks the feed depth setting status for each numbered path, screens out the unset items and classifies them into the adjustment path number, merges the numbers and reorders them, and obtains the adjustable feed operation path segment sequence; The load matching module locates the cross coordinates of the nodes in the path segment based on the path segments included in the sequence of adjustable feed operation path segments, delimits the spatial area according to the node coordinates and maps them to the spatial distribution grid, extracts the node quantity data in the area, compares the density benchmark value to select the area number whose node density reaches the benchmark value, and associates the corresponding path segment number to obtain the staggered operation path segment registration table; The segment analysis module extracts the direction vector sequence of continuous path points in the segment based on the path segments in the staggered operation path segment registration table, identifies the line segments with continuous direction changes, analyzes the consistency of the point arrangement sequence, and selects the path segment numbers with both direction consistency and point continuity to obtain the continuous processing segment group; The posture adjustment module calls the end effector posture value of the path segment in the continuous processing segment group, extracts the carving surface direction of the path segment, compares the posture angle relationship between the two, screens out the path segments whose offset exceeds the reference angle, and inserts the corrected segment into the original processing trajectory after correcting the posture angle to obtain the special-shaped stone processing path set.

[0041] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for processing three-dimensional special-shaped stone by a robot, characterized in that: The following steps are involved: S1: Obtain the robot engraving path segment, extract the particle image in the path area and analyze the arrangement state, divide the density category according to the particle spacing, and generate a particle arrangement difference segment identification set; S2: Based on the path segments marked by the particle arrangement difference segment identification set, the path segments are classified according to the particle distribution state, the tool load range is called, the path segments are mapped and matched with the load conditions, the path index is extracted, and the path segments without the feed depth set are classified into the adjustment range to obtain the adjustable feed operation path segment sequence; S3: Based on the path segments in the adjustable feed operation path segment sequence, the path segment node coordinates are mapped to the spatial distribution grid, the number of grid area nodes is counted, and a staggered operation path segment registration table is obtained; S4: based on the path line group in the staggered operation path segment registration table, extract the path point direction vector and verify the direction continuity, analyze the path continuity in combination with the connection sequence, and obtain the continuous processing segment group; S5: Based on the path direction of the segments in the continuous processing segment group, the difference between the posture of the end effector of each path segment and the direction of the engraving surface is analyzed, and the angle of the deviation segment is adjusted and then connected to the trajectory to obtain a special-shaped stone processing path set.

2. The method for processing three-dimensional special-shaped stone by a robot according to claim 1, characterized in that: The particle arrangement difference segment identification set includes particle arrangement direction data, average particle spacing interval, dense and sparse area boundary marks, the adjustable feed operation path segment sequence includes a path segment number set, a load adaptation annotation item, and a feed depth setting identifier, the staggered operation path segment registration table includes the intersection node space coordinates, node density distribution indicators, and staggered path segment index numbers, the continuous processing segment group includes direction continuous path segments, node sequence mapping relationships, and stable continuation segment numbers, and the special-shaped stone processing path set includes posture adjustment path segments, processing trajectory sequence indexes, and end effector angle correction values.

3. The method for processing three-dimensional special-shaped stone by a robot according to claim 1, characterized in that: The specific steps are: S101: Obtain the robot path segment in the stone carving task, extract the stone surface image in the path corresponding area, identify the particle arrangement direction and particle spacing distribution, extract the particle position and arrangement characteristics in the path segment, and obtain the particle arrangement recognition result; S102: Based on the particle arrangement recognition result, analyzing the connectivity of the particle directions in the path segment, combining the particle spacing variation trend and the quantity distribution range, classifying the path segment according to a preset classification standard, and generating a particle density segment classification result; S103: calling the particle density segment classification result, marking the path points of the difference boundaries in the path segment, updating the density distribution state in the path index and structure information, and obtaining a particle arrangement difference segment identification set.

4. The method for processing three-dimensional special-shaped stone by a robot according to claim 1, characterized in that: The specific steps are: S201: Based on the path segments marked in the particle arrangement difference segment identification set, the path segments are divided according to the continuity of the particle distribution state and the density distribution trend in the path segments, and the path segments are classified into differential category identifications according to the particle number distribution value and the spacing mean, so as to obtain a path segment classification label set; S202: Based on the path segment classification mark set, read the load range value set on the tool structure, extract the total length of each type of path segment and correspond it to the category number, calculate the total execution length ratio of the category path segment, filter the matching path segment numbers, and obtain the executable path segment index table; S203: calling the executable path segment index table, detecting the record items in the path number that do not have the feed depth set, screening the number and merging and marking it with the original index, re-arranging the data index according to the merged path set, and obtaining an adjustable feed operation path segment sequence.

5. The method for processing three-dimensional special-shaped stone by a robot according to claim 4, characterized in that: The calculation formula of the total execution length ratio of the category path segment is specifically: ； in, Represents the total execution length ratio of the path segment of category i, represents the total execution length of the i-th path segment, represents the arithmetic mean of the load values ​​set for all path segments in the i-th path segment, represents the median of the load values ​​set for all path segments in the i-th path segment. represents the load weight coefficient of the structure corresponding to the i-th path segment, Represents the arithmetic mean of the total execution length of all class path segments, Represents the total execution length of the k-th path segment, Represents the average value of the load values ​​set for all path segments in the k-th path segment. represents the structural load weight of the k-th path segment, and K represents the total number of path segment categories.

6. The method for processing three-dimensional special-shaped stone by a robot according to claim 1, characterized in that: The specific steps are: S301: based on the path segments in the adjustable feed operation path segment sequence, extracting node coordinate data in the path segment, dividing the node position into a spatial distribution grid according to the coordinate range, identifying the grid number to which the node belongs and establishing a mapping index, and obtaining a node grid belonging relationship table; S302: Based on the node grid attribution relationship table, the total number of nodes associated with the grid number and the corresponding coordinate boundary are extracted, the node spatial distribution density value within the unit area under the number is calculated, and the grid number whose distribution density exceeds the reference value is selected in combination with the density reference standard to generate a high-density node grid number set; S303: calling the high-density node grid number set, filtering the path segment index information associated with the number set, merging the corresponding path segments into group data according to the grid numbers, and adding interleaved area marks to obtain an interleaved operation path segment registration table.

7. The method for processing three-dimensional special-shaped stone by a robot according to claim 6, characterized in that: The specific calculation formula for the node spatial distribution density value per unit area under the number is: ； in, Represents the node spatial distribution density value per unit area under the grid number z. represents the total number of nodes associated with the grid number z, represents the qth weighting factor associated with the node distribution, Represents the relative spatial difference between the grid number z and the qth node, Represents a coordinate value of the sth node in the grid number z, Represents the average value of the coordinates of all nodes in the grid number z. represents the boundary area corresponding to the grid number z, Representative The coefficient of the boundary influence factor of the term, Represents the grid number z. The boundary morphology adjustment parameters are u, the total number of factors involved in the weighted calculation of node space difference, v is the total number of nodes contained in the grid number z, and w is the number of types of factors affecting the grid boundary.

8. The method for processing three-dimensional special-shaped stone by a robot according to claim 1, characterized in that: S401: based on the path line groups in the staggered operation path segment registration table, extract the direction vector information of adjacent points in each path, sort the vector arrangement order according to the path node order, analyze the consistency of the direction change trend between adjacent vectors, and obtain the path direction continuous segment group; S402: Based on the path direction continuous segment group, extract the arrangement order of the corresponding path points in the paragraph, detect whether there is a sequence interruption or reverse jump of the path points, and select the path segment information that maintains a continuous point sequence to obtain a path point sequence continuous segment group; S403: calling the path point sequential continuous segment group, performing matching processing against the existing directional continuity information, verifying the path line group that satisfies both directional stability and point arrangement continuity, and obtaining a continuous processing segment group.

9. The method for processing three-dimensional special-shaped stone by a robot according to claim 1, characterized in that: The specific steps are: S501: based on the path direction of the path segments in the continuous processing segment group, extract the end effector posture data, and compare it with the corresponding engraving plane direction, identify the path segments whose spatial direction difference reaches the posture deviation reference value, and obtain the posture deviation meeting the condition segment index set; S502: calling the index set of the segment whose posture deviation meets the condition, reading the end angle information of the corresponding path segment, adjusting the posture angle according to the preset rule, and binding the adjustment value to the path segment posture information to generate a posture correction path segment set; S503: Based on the posture correction path segment set, the correction segment is added to the processing trajectory data, the path segment number and the sequence index are updated, and the processing trajectory is adjusted to obtain the special-shaped stone processing path set.

10. A robot system for processing three-dimensional special-shaped stones, characterized in that: The method for processing three-dimensional special-shaped stone by a robot according to any one of claims 1 to 9 comprises the following steps: The path construction module obtains the robot processing path segment of three-dimensional special-shaped stone carving, collects the particle image of the stone surface in the area covered by the path segment, partitions the particle arrangement direction, spacing distribution and boundary changes in the image, delineates the structural distribution inside the path segment according to the particle concentration degree, and generates a particle arrangement difference area identification set; The structure recognition module identifies the particle quantity distribution and arrangement spacing of the concentrated path segment based on the particle arrangement difference area, identifies the particle dense, moderate and sparse segments, reads the load-bearing interval marked in the tool structure equipped with the robot, matches the particle category and load condition to which the path segment belongs, extracts the path segment number that meets the processing requirements, checks the feed depth setting status for each numbered path, screens out unset items and classifies them into the adjustment path number, merges the numbers and reorders them, and obtains an adjustable feed operation path segment sequence; The load matching module locates the intersection coordinates of the nodes in the path segment based on the path segments included in the adjustable feed operation path segment sequence, delimits the spatial area according to the node coordinates and maps them to the spatial distribution grid, extracts the node quantity data in the area, compares the density benchmark value to select the area number whose node density reaches the benchmark value, and associates the corresponding path segment number to obtain the staggered operation path segment registration table; The segment analysis module extracts the direction vector sequence of continuous path points in the segment based on the path segments in the staggered operation path segment registration table, identifies the line segments with continuous direction changes, analyzes the consistency of the point arrangement sequence, and selects the path segment numbers with both direction consistency and point continuity to obtain a continuous processing segment group; The posture adjustment module calls the end effector posture value of the path segment in the continuous processing segment group, extracts the carving surface direction of the path segment, compares the posture angle relationship between the two, screens out the path segment whose offset exceeds the reference angle, corrects the posture angle and inserts the corrected segment into the original processing trajectory to obtain the special-shaped stone processing path set.

Images