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

Abstract

The present invention relates to the field of stone processing technology, specifically to a method and system for robot processing three-dimensional special-shaped stone, comprising the following steps: obtaining path segments and extracting particle images, dividing density categories to identify segments, classifying path segments to match load conditions and screening out indexes, mapping node coordinates to generate an interlaced segment table, extracting direction vectors and sequence information to identify continuous segments, adjusting posture angles to access trajectories, and generating a processing path set. In the present invention, by extracting particle images and dividing density categories, identifying particle distribution differences, combining spatial distribution grids to identify interlaced node areas, defining structural segments, jointly checking the consistency of path segment directions and sequential coherence, controlling path interruptions and jumps, analyzing the difference between the end posture and the carving surface direction and coordinating the angle adjustment, reducing cutting offset and contour misalignment caused by posture mismatch during processing, and enhancing path stability and carving coherence during the three-dimensional special-shaped stone carving process.

Description

Technical Field

[0001] The present 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 encompasses physical processing processes such as cutting, carving, grinding, drilling, and forming for natural and artificial stone. This field covers the manufacturing process of stone products for use in architectural decoration, arts and crafts, and landscaping. The core content of this technical field lies in the precise processing of stone in terms of size, shape, and texture through various types of mechanical processing equipment or automatic control systems. The development direction of stone processing technology includes the diversification of processing techniques, the intelligence of mechanical equipment, and the improvement of processing accuracy. In particular, CNC equipment, multi-axis robotic arms, and programmable control systems have been widely used to achieve efficient processing in the expression of carved details and complex three-dimensional structures. This field is closely related to automation technology and mechanical structure design, especially when it comes to the complex forming requirements of irregular curved surfaces and large-scale stone workpieces.

[0003] Among them, the method of robot processing of three-dimensional special-shaped stones refers to a method of using a multi-axis linkage industrial robot combined with digital modeling and path planning to perform three-dimensional processing on natural stones with large volume and complex structure. The subject of this patent mainly targets technical issues 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 processing, semi-finishing and finishing of special-shaped stones are completed. The means adopted 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 shaping of the stone.

[0004] Existing machining path control technologies primarily rely on direct analysis of CAD model paths and the issuance of CNC commands. These technologies lack real-time recognition and adaptation to the actual surface characteristics and structural density variations of the stone. This makes it impossible to dynamically respond to uneven particle distribution and drastic local density variations on the stone surface through the machining path. This often leads to excessive or insufficient tool load in localized areas, resulting in edge chipping and overcutting. For example, when processing areas with high-density mineral concentrations, maintaining a constant feed depth can easily cause tool vibration and damage to the engraved surface. Furthermore, existing path control strategies often utilize a fixed step-over distance and lack spatially distributed staggered avoidance strategies between path segments. This can easily lead to path overlap and posture misalignment on complex three-dimensional structures, resulting in reduced machining continuity and poor surface consistency. The end-effector posture is typically set only during the initial modeling phase, lacking real-time feedback and adjustment for posture deviations at the end of the path. This makes it impossible to ensure directional consistency between the machining trajectory and the target engraving surface, and can easily cause local relief offset or ghosting in curved transition areas. The path control logic of existing machining technologies remains model-centric, ignoring the variability of actual stone structures during machining. 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:

[0006] 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:

[0007] S1: Obtain the robot engraving path segment, extract the particle image in the path area and analyze the arrangement state, divide the density categories according to the particle spacing, and generate a particle arrangement difference segment identification set;

[0008] 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 set feed depth are classified into the adjustment range to obtain the adjustable feed operation path segment sequence;

[0009] S3: Based on the path segments in the adjustable feed operation path segment sequence, the path segment node coordinates are mapped to a spatial distribution grid, the number of grid area nodes is counted, and a staggered operation path segment registration table is obtained;

[0010] 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;

[0011] 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 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.

[0012] 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 markers, 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 index, and staggered path segment index number, the continuous processing segment group includes direction continuous path segments, node sequence mapping relationship, and stable continuation segment number, and the special-shaped stone processing path set includes posture adjustment path segments, processing trajectory sequence index, and end effector angle correction value.

[0013] As a further solution of the present invention, the specific steps of S1 are:

[0014] 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 within the path segment, and obtain the particle arrangement recognition result;

[0015] S102: Based on the particle arrangement recognition result, analyzing the connectivity of the particle directions within the path segment, combining the particle spacing variation trend and quantity distribution range, classifying the path segment according to a preset classification standard, and generating a particle density segment classification result;

[0016] 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.

[0017] As a further solution of the present invention, the specific steps of S2 are:

[0018] S201: Based on the path segments marked in the particle arrangement difference segment identification set, the path segments are divided according to the continuity and density distribution trend of the particle distribution state in the path segments, and the path segments are classified into differential category identifications according to the particle number distribution value and the mean spacing value to obtain a path segment classification label set;

[0019] 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 with the category number to which it belongs, calculate the total execution length ratio of the category path segment, screen the matching path segment numbers, and obtain an executable path segment index table;

[0020] S203: calling the executable path segment index table, detecting the record items in the path number that do not have the feed depth set, filtering 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.

[0021] As a further solution of the present invention, the calculation formula for the total execution length ratio of the category path segment is specifically:

[0022] ;

[0023] in, Represents the proportion of the total execution length 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 category 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.

[0024] As a further solution of the present invention, the specific steps of S3 are:

[0025] S301: Based on the path segments in the adjustable feed operation path sequence, node coordinate data within the path segments are extracted, node positions are divided into spatial distribution grids according to the coordinate range, the grid numbers to which the nodes belong are identified and a mapping index is established to obtain a node-grid affiliation relationship table;

[0026] 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 spatial distribution density of the nodes per unit area under the number is calculated, and the grid numbers with distribution density exceeding the reference value are screened in combination with the density reference standard to generate a high-density node grid number set;

[0027] 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 number, and adding an interleaved area mark to obtain an interleaved operation path segment registration table.

[0028] As a further solution of the present invention, the calculation formula of the node spatial distribution density value per unit area under the numbering is specifically:

[0029] ;

[0030] in, Represents the node spatial distribution density value per unit area under the grid number z number, 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, Represents the grid number z boundary morphology adjustment parameters, u is the total number of factors involved in the weighted calculation of node spatial difference, v is the total number of nodes contained in grid number z, and w is the number of types of factors affecting the grid boundary.

[0031] As a further solution of the present invention, the specific steps of S4 are:

[0032] 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, arrange the vector arrangement order according to the path node sequence, analyze the consistency of the direction change trend between adjacent vectors, and obtain a path direction continuous segment group;

[0033] 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 path segment information that maintains a continuous point sequence to obtain a path point sequence continuous segment group;

[0034] 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 directional stability and point arrangement continuity, and obtaining a continuous processing segment group.

[0035] As a further solution of the present invention, the specific steps of S5 are:

[0036] S501: Based on the path directions of the path segments in the continuous processing segment group, extract the end effector posture data, compare it with the corresponding engraving plane direction, identify the path segments whose spatial direction differences reach the posture deviation reference value, and obtain the index set of the segments that meet the posture deviation conditions;

[0037] S502: calling the index set of segments whose posture deviations meet the conditions, reading the end angle information of the corresponding path segment, adjusting the posture angle according to a preset rule, and binding the adjustment value to the path segment posture information to generate a posture-corrected path segment set;

[0038] 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.

[0039] The robot system for processing three-dimensional special-shaped stones includes:

[0040] The path construction module obtains the robot processing path segments for three-dimensional special-shaped stone carving, collects particle images of the stone surface within the area covered by the path segments, partitions the image based on the particle arrangement direction, spacing distribution, and boundary changes, delineates the structural distribution within the path segments according to the degree of particle concentration, and generates a set of identification sets for areas with different particle arrangement.

[0041] The structure recognition module identifies the particle quantity distribution and arrangement spacing of the concentrated path segments 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 carried by the robot, matches the particle category to the path segment with the load condition, extracts the path segment number that meets the processing requirements, checks the feed depth setting status of each numbered path, filters out unset items and assigns them to the adjustment path number, merges the numbers and reorders them to obtain a sequence of adjustable feed operation path segments;

[0042] The load matching module locates the intersection coordinates of nodes in the path segments based on the path segments included in the adjustable feed operation path segment sequence, demarcates a spatial region according to the node coordinates and maps the spatial distribution grid, extracts node quantity data in the region, compares the density benchmark value to select the region number whose node density reaches the benchmark value, associates the corresponding path segment number, and obtains a staggered operation path segment registration table;

[0043] 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 line segments with continuous direction changes, analyzes the consistency of the point arrangement sequence, and selects the path segment numbers that have both directional consistency and point continuity to obtain a continuous processing segment group;

[0044] 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.

[0045] Compared with the prior art, the advantages and positive effects of the present invention are:

[0046] 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 differences between the end posture and the carving surface direction in conjunction with 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

[0047] 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 paying any creative work.

[0048] Figure 1 Schematic diagram of the steps of the present invention;

[0049] Figure 2 It is a system module diagram of the present invention. DETAILED DESCRIPTION

[0050] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0051] 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 an "exemplary" in the present invention should not be interpreted as being preferred or advantageous over other embodiments or designs. Rather, the use of the word "exemplary" is intended to present concepts in a concrete manner. Furthermore, in the embodiments of the present invention, "and / or" can mean both or either of the two.

[0052] In the embodiments of the present invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, when the distinction is not emphasized, the meanings they convey are the same. The terms "of," "corresponding," and "corresponding" may sometimes be used interchangeably. It should be noted that, when the distinction is not emphasized, the meanings they convey are the same.

[0053] In the embodiments of the present invention, sometimes a subscript such as W1 may be written as a non-subscript such as W1. When the difference is not emphasized, the meanings to be expressed are the same.

[0054] In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0055] See also Figure 1 The method for processing three-dimensional special-shaped stone by a robot comprises the following steps:

[0056] S1: Obtain the robot's preset path segments in the stone carving task, extract the surface particle image of the stone corresponding to the path in the robot's operating section, analyze the particle arrangement direction, average particle spacing, and changes in the boundaries between dense and sparse areas, divide the path segments into density categories based on the average particle spacing and particle number, group and mark the partition boundary points, and update the path data index to generate a particle arrangement difference segment identification set;

[0057] 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 status to which they 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 based on the matching results. The path segments that have not completed the feed depth setting are simultaneously included in the adjustment range to obtain the adjustable feed operation path segment sequence;

[0058] S3: Based on the path segments in the adjustable feed operation path 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 cell is counted. The grid number sequence is called and compared with the node density data. The cells with a number exceeding the density threshold are marked as interleaved segment areas. The corresponding path segment numbers are summarized to obtain the interleaved operation path segment registration table.

[0059] 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. The segments with consistent directions are identified based on the direction continuity between the path points. The lengths of the line segments between the connection points are compared with the sequence of the path points to identify the set of paths with stable directions and continuous points, thus obtaining the continuous processing segment group.

[0060] 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.

[0061] 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 identifiers, 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.

[0062] The specific steps of S1 are:

[0063] 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 within the path segment, and obtain the particle arrangement recognition result;

[0064] 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. 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. 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 lattice with grayscale values in the range of 65 to 180 in the image is extracted. Then, the contour tracking operation is performed on the particle edge 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 collected image coordinate system. The particle center point data within the coverage range of the path segment is screened as the particle set in the current path segment. On this basis, the connection direction between the particle center points is calculated, the pairwise connection direction vector is defined, and the angle between them is counted. 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, so the main arrangement direction is determined to be 10 degrees. Then, 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. Then, 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 the partition boundary points. The position index number of the path segment in the path set is marked, and the particle arrangement direction, average particle spacing and number of particles are returned as record items. The structural attribute field of the path segment is updated to obtain the particle arrangement recognition result.

[0065] S102: Based on the particle arrangement recognition results, the connectivity of the particle directions within the path segment is analyzed. In combination with the particle spacing variation trend and quantity distribution range, the path segment is classified according to a preset classification standard to generate a particle density segment classification result.

[0066] First, the main arrangement direction value sequence identified in each path segment is extracted, and the particle arrangement angle variation range is divided according to the arrangement continuity of the 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 adjacent angle difference is greater than 10°, such as 17°, 28°, and 39°, it is recorded as a region with poor directional connectivity, and the region with good directional connectivity is marked as a stable segment. Secondly, the particle spacing values of each path segment are arranged in sequence to construct a change 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 spacing trend stable segment. If a segment's change trend shows a regular increase or decrease, it is considered to be a spacing trend fluctuation. The particle spacing trend type is recorded in units of path segments. 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 is more than 6, it is defined as a high-density segment, and if it 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. Path segments with two of the three conditions meeting the stability standard are classified as consistent density segments. If two of the three conditions fluctuate, they are classified as density fluctuation segments. The division results of various path segments are accompanied by the original path segment number and boundary coordinates to generate the particle density segment classification results.

[0067] 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;

[0068] 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 end point of each path segment in the processing path, extract the path points at the intersection 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 intersection between the second 2 and the first 3 as the basis for identifying the difference boundary path point. After completing the boundary path point extraction, press The original path index list is updated in order of 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 association key value, and the density interval number obtained in the particle density segment classification is additionally 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 number P45 is category 3, and the boundary point position is between the 27th and 28th points. Then the new 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.

[0069] The specific steps of S2 are:

[0070] S201: Based on the path segments marked in the particle arrangement difference segment identification set, the path segments are divided according to the continuity and density distribution trend of the particle distribution state in the path segments, and the path segments are classified into differential category identifications according to the particle number distribution value and the mean spacing value to obtain the path segment classification label set;

[0071] 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 according to the length direction of the path segment 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. The continuity is evaluated according to the proportion of fluctuation times and the span of the variation interval. If the spacing difference in the continuous segment exceeds 40% of the maximum value of the total segment spacing, it is recorded as a discontinuous state, otherwise it is defined as a continuous distribution state. For example, the path number P12, the adjacent particle spacings 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 analysis, 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 obtained according to the definition of "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, a comparison method is used to set the low, medium and high density division intervals. 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 identifier 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 tag set.

[0072] 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 with the category number to which it belongs, calculate the total execution length ratio of the category path segment, screen the matching path segment numbers, and obtain an executable path segment index table;

[0073] The formula for calculating the proportion of the total execution length of the category path segment is as follows:

[0074] ;

[0075] in, Represents the proportion of the total execution length 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 category 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;

[0076] Assume the following values:

[0077] =300 meters, =1.5 times, =1.4 times, =0.8, =300 meters, =300 meters, =1.5 times, =0.8, K=5;

[0078] Example of calculation derivation process:

[0079] calculate ;

[0080] ;

[0081] ;

[0082] calculate ;

[0083] ;

[0084] =1599;

[0085] Calculate the final result :

[0086] ;

[0087] Result interpretation:

[0088] The results show that the total execution length of the i-th path segment accounts for 5.31%, indicating that its execution resource share among all path segments is only 5.31%. This value is determined by the path segment length, structural load weight, and load variation, while the denominator reflects the weighted total execution volume of all 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, the path segment is given a lower execution priority in task allocation.

[0089] S203: calling the executable path segment index table, detecting the record items in the path number that do not have the feed depth set, filtering the numbers and merging and marking them with the original indexes, and re-arranging the data indexes according to the merged path set to obtain the adjustable feed operation path segment sequence;

[0090] 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 effective feed depth has not been set for the numbered path segment. 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 is spliced and attached to the newly marked field "Depth Status: Unset" to form data with a complete structural description. Then, based on the path number sequence, all data rows are rearranged in ascending order. 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 appended as the data source of the feed parameter adjustment list. For example, after rearrangement, the corresponding indexes of numbers P102, P105, and P108 are 12, 17, and 23 respectively, then the new record in the rearranged 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.

[0091] The specific steps of S3 are:

[0092] S301: Based on the path segments in the adjustable feed operation path sequence, node coordinate data within the path segments are extracted, node positions are divided into spatial distribution grids according to the coordinate range, the grid numbers to which the nodes belong are identified, and a mapping index is established to obtain a node-grid affiliation relationship table;

[0093] First, all node coordinate information in each path segment is extracted, the spatial position of the node is traversed, and the coordinate data is parsed into a set of X, Y, and Z axis coordinate values in three-dimensional space. Then, the spatial grid boundary is set according to the working range of the stone processing platform, and the minimum segmentation unit within the three-dimensional range is delineated. Each unit volume is 10mm×10mm×10mm and is assigned a unique number. A position matching operation is performed on each node coordinate, and the X, Y, and Z coordinate values of each node are compared with the grid boundary respectively to determine the spatial grid number interval in which the coordinate point falls. After matching, a one-to-one mapping relationship is formed 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 10mm unit grid condition, 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, the node grid ownership relationship table is obtained.

[0094] 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 spatial distribution density of the nodes per unit area under the number is calculated, and the grid numbers with distribution density exceeding the reference value are screened based on the density reference standard to generate a high-density node grid number set;

[0095] The calculation formula for the node spatial distribution density value per unit area under the number is as follows:

[0096] ;

[0097] in, Represents the node spatial distribution density value per unit area under the grid number z number, 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, Represents the grid number z The boundary morphology adjustment parameters are u, the total number of factors involved in the weighted calculation of node spatial 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.

[0098] This formula is used to calculate the spatial distribution density of nodes within a specific area. The following is a detailed explanation of each part of the formula and the calculation derivation:

[0099] Assume the following values: = node, =0.8, =1.2, = 2.5 meters, =3.0 m, =50.0 m, =52.5 meters, =51.25 meters, =100 square meters, =1.5, =1.0, =0.3 =0.2;

[0100] Formula calculation example:

[0101] Consider the case of u=2, v=2, w=2, the sum of weighted node space differences:

[0102] ;

[0103] Mean square error of node coordinate values:

[0104] ;

[0105] The square root is:

[0106] ;

[0107] Boundary Adjustment Sum:

[0108] ;

[0109] Substituting these values into the density formula:

[0110] ;

[0111] This result indicates that the calculated The spatial distribution density of nodes per unit area in grid number z is 1.225, indicating that the nodes are relatively dense in this area. This reflects the dense distribution of nodes in the grid area, which can be used to further determine the optimization or enhancement measures of network resources.

[0112] 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 by grid number, and adding an interleaved area mark to obtain an interleaved operation path segment registration table;

[0113] 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, so 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 Flag: Yes" is added. Otherwise, the mark "Interlaced Flag: 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 by the path segment number. Each record is accompanied by the path segment number, number of nodes, grid number distribution and interlaced flag field value. For example, the output corresponding to the number P204 is "P204 | 6 nodes | G12-04, G12-05, G13-04 | Interlaced Flag: Yes". Finally, all marking results are summarized to generate the interlaced operation path segment registration table.

[0114] The specific steps of S4 are:

[0115] 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, arrange the vector arrangement order according to the path node sequence, analyze the consistency of the direction change trend between adjacent vectors, and obtain the path direction continuous segment group;

[0116] 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 difference between the three-dimensional coordinates of the next node and the previous node. The complete vector sequence within the path segment is sequentially constructed. Then, the direction change trend of the vector sequence is analyzed. 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° and 17°, it can be identified that the first three segments constitute a directional continuous segment. 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, end node, 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 a continuous segment and 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.

[0117] 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 the path point sequence continuous segment group;

[0118] 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 index values of adjacent nodes in the node number sequence is judged. If the index difference between two adjacent node numbers is 1, it is considered to be sequentially continuous, otherwise it is marked as a sequence interruption. For example, in the segment numbered P308, the node sequence is N12, N13, N15, and N16, then a jump between N13 and N15 is identified, and the interruption position is recorded. If a reverse order occurs, such as the node number sequence is N25, N24, and N23, the continuity is broken. During the judgment, it is identified that the index value shows a decreasing trend and is classified as a reverse jump type. The path segment is screened according to the integrity of the node number sequence within each path segment. All segments with increasing indexes and no jumps are retained as valid sequence segments, and grouped and sorted according to the path number. The path number, segment start and end node numbers, and continuity judgment status are recorded in the data structure. For the example numbered P308, only N15 to N16 constitute a continuous segment. Finally, it is included and output in the format of "P308 | N15-N16 | Sequence status: continuous". All path segment segments that meet the sequence integrity conditions are summarized to obtain a path point sequence continuous segment group.

[0119] 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 direction stability and point arrangement continuity, and obtaining a continuous processing segment group;

[0120] 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 3D 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 number, 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 a dot product operation on each directional vector in the calling segment 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 approximately 2.86°, s_i≈0.9987, which meets the directional continuity requirement. To further verify the path line group of directional stability and point arrangement continuity, all consecutive paragraphs need to be combined into a path line group 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 start and end points of the paragraph is less than δ (set the paragraph continuity threshold, for example, 5mm), it is determined to meet the continuity requirement, and finally a continuously processed paragraph group is obtained.

[0121] The specific steps of S5 are:

[0122] S501: Based on the path directions of the path segments in the continuous processing segment group, the end effector posture data is extracted and compared with the corresponding engraving plane direction, and the path segments whose spatial direction differences reach the posture deviation reference value are identified, and the index set of the segments with posture deviation meeting the conditions is obtained;

[0123] 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 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 node engraving surface 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 range, the current path segment number and the corresponding node index number can be recorded. In the path segment numbered P218, there are nodes N01 to N08, among which nodes N03 and N04 have an angular offset in the Z-axis direction exceeding the attitude deviation reference value of ±10 degrees. Then, numbered P218 is marked as a segment with attitude deviation, and N03-N04 is recorded as a deviation key node interval. 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. The path number, start and end node number, deviation type, offset direction and other information are included in the structured list and output. All path segments that meet the attitude deviation reference requirements are summarized to obtain the attitude deviation qualified segment index set.

[0124] S502: Calling the index set of segments whose posture deviations meet the conditions, reading the end angle information of the corresponding path segment, adjusting the posture angle according to a preset rule, and binding the adjustment value to the path segment posture information to generate a posture-corrected path segment set;

[0125] 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 as ±5 degrees, compare the Z axis offset of 15 degrees with the offset reference value, the difference of 10 degrees 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. , record the Z-axis correction value as 5 degrees, insert the Z-axis angle correction field into the node record structure, and 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 and output according to the original numbers to generate a posture correction path segment set.

[0126] S503: Based on the posture correction path segment set, the correction segment is added to the processing trajectory data, the path segment number and sequence index are updated, and the processing trajectory is adjusted to obtain the special-shaped stone processing path set;

[0127] First, the path segment structure content is read from the original processing trajectory dataset, and 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 using a sequential insertion method and replaced with the original segment data. The original segment start and end point numbers are retained and the numbers are synchronously updated to P042-A. The path index field is reassigned, and the subsequent path segment numbers are updated in sequence from P043 to P048. At the same time, the segment number mapping table recorded in the trajectory structure is adjusted synchronously. 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. After confirming that there is no conflict in the segment number, the trajectory index table is written to the trajectory index summary table. The path segment sequence is traversed and checked. After confirming that the number increment logic 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. The complete trajectory set data containing the corrected posture segments, updated number sequence, and adjusted trajectory structure is output to obtain the special-shaped stone processing path set.

[0128] See also Figure 2 , a robot system for processing three-dimensional special-shaped stones, including:

[0129] The path construction module obtains the robot processing path segments for three-dimensional special-shaped stone carving, collects particle images of the stone surface within the area covered by the path segments, partitions the image based on the particle arrangement direction, spacing distribution, and boundary changes, delineates the structural distribution within the path segments according to the degree of particle concentration, and generates a set of identification sets for areas with different particle arrangement.

[0130] The structure recognition module identifies the particle number distribution and arrangement spacing of concentrated path segments based on the particle arrangement difference area, identifies dense, moderate, and sparse particle segments, reads the load range marked in the tool structure carried by the robot, matches the particle type to the load condition, extracts the path segment number that meets the processing requirements, checks the feed depth setting status of each numbered path, filters out unset items and assigns them to the adjustment path number, merges the numbers and reorders them to obtain a sequence of adjustable feed operation path segments;

[0131] The load matching module locates the intersection coordinates of nodes in the path segments based on the path segments included in the adjustable feed operation path segment sequence, demarcates spatial regions according to the node coordinates and maps them to the spatial distribution grid, extracts the node quantity data within the region, compares the density benchmark value to select the region numbers with node density reaching the benchmark value, and associates the corresponding path segment numbers to obtain the staggered operation path segment registration table;

[0132] 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 line segments with continuous direction changes, analyzes the consistency of the point arrangement sequence, and selects the path segment numbers that have both directional coherence and point continuity to obtain the continuous processing segment group;

[0133] 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, corrects the posture angle, and inserts the corrected segment into the original processing trajectory to obtain the special-shaped stone processing path set.

[0134] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A method for processing three-dimensional special-shaped stones 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 categories 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 set feed depth 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 a 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 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 path segment number set, load adaptation annotation item, feed depth setting identifier, the staggered operation path segment registration table includes intersection node spatial coordinates, node density distribution index, staggered path segment index number, the continuous processing segment group includes direction continuous path segment, node sequence mapping relationship, stable continuation segment number, the special-shaped stone processing path set includes posture adjustment path segment, processing trajectory sequence index, and end effector angle correction value.

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 within 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 within the path segment, combining the particle spacing variation trend and 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 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.

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 and density distribution trend of the particle distribution state in the path segments, and the path segments are classified into differential category identifications according to the particle number distribution value and the mean spacing value 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 with the category number to which it belongs, calculate the total execution length ratio of the category path segment, screen the matching path segment numbers, and obtain an 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, filtering 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 for the total execution length ratio of the category path segment is as follows: ； in, Represents the proportion of the total execution length 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 category 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 sequence, node coordinate data within the path segments are extracted, node positions are divided into spatial distribution grids according to the coordinate range, the grid numbers to which the nodes belong are identified and a mapping index is established to obtain a node-grid affiliation 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 spatial distribution density of the nodes per unit area under the number is calculated, and the grid numbers with distribution density exceeding the reference value are screened 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 number, and adding an interleaved area mark 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 calculation formula for 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 number, 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, Represents the grid number z boundary morphology adjustment parameters, u is the total number of factors involved in the weighted calculation of node spatial difference, v is the total number of nodes contained in 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, arrange the vector arrangement order according to the path node sequence, analyze the consistency of the direction change trend between adjacent vectors, and obtain a 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 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 direction 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 directions of the path segments in the continuous processing segment group, extract the end effector posture data, compare it with the corresponding engraving plane direction, identify the path segments whose spatial direction differences reach the posture deviation reference value, and obtain the index set of the segments that meet the posture deviation conditions; S502: calling the index set of segments whose posture deviations meet the conditions, reading the end angle information of the corresponding path segment, adjusting the posture angle according to a preset rule, and binding the adjustment value to the path segment posture information to generate a posture-corrected 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 segments for three-dimensional special-shaped stone carving, collects particle images of the stone surface within the area covered by the path segments, partitions the image based on the particle arrangement direction, spacing distribution, and boundary changes, delineates the structural distribution within the path segments according to the degree of particle concentration, and generates a set of identification sets for areas with different particle arrangement. The structure recognition module identifies the particle quantity distribution and arrangement spacing of the concentrated path segments 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 carried by the robot, matches the particle category to the path segment with the load condition, extracts the path segment number that meets the processing requirements, checks the feed depth setting status of each numbered path, filters out unset items and assigns them to the adjustment path number, merges the numbers and reorders them to obtain a sequence of adjustable feed operation path segments; The load matching module locates the intersection coordinates of nodes in the path segments based on the path segments included in the adjustable feed operation path segment sequence, demarcates a spatial region according to the node coordinates and maps the spatial distribution grid, extracts node quantity data in the region, compares the density benchmark value to select the region number whose node density reaches the benchmark value, associates the corresponding path segment number, and obtains a 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 line segments with continuous direction changes, analyzes the consistency of the point arrangement sequence, and selects the path segment numbers that have both directional 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