A laser intelligent cleaning process composite path planning method and system

Abstract

The application provides a laser intelligent cleaning process composite path planning method and system, relates to the trajectory planning technology and visual image processing and analysis technology fields in the field of artificial intelligence, and relates to the laser intelligent cleaning process composite path planning method, which comprises the following steps: acquiring a binary image corresponding to a workpiece to be cleaned, and performing regional division on the binary image to obtain a plurality of sub-partitions; performing path planning on each sub-partition according to a set path heuristic mode to obtain path planning data of the corresponding sub-partition; and performing path planning on all sub-partitions based on an improved ant colony algorithm according to the path planning data of each sub-partition to obtain global composite path planning. Through the combination of multiple path planning technologies and the global path access link (global path planning) combined with the improved ant colony algorithm, the global series connection coverage of all sub-partitions is realized, the laser cleaning path can be accurately and efficiently planned, and the cleaning accuracy and efficiency are improved.

Description

Technical Field

[0001] This invention relates to trajectory planning technology and visual image processing and analysis technology in the field of artificial intelligence, and more specifically, to a composite path planning method and system for a laser intelligent cleaning process. Background Technology

[0002] Surface cleaning plays an important role in quality maintenance and performance assurance in modern industrial manufacturing. Among the many cleaning methods (such as mechanical cleaning, chemical cleaning, ultrasonic cleaning, and laser cleaning), laser cleaning is considered to be the most promising green cleaning technology at present.

[0003] With the rapid development of laser technology, laser cleaning can utilize different removal principles such as thermal ablation, resonance, and thermal expansion to achieve the cleaning process of workpiece surfaces. Due to its advantages such as high controllability, environmental friendliness, non-contact, and non-destructive properties, its application areas are constantly expanding.

[0004] However, existing laser cleaning technologies rely more on manual calibration and full-coverage scanning, and the accuracy of laser cleaning path planning is relatively low. For larger cleaning targets or paths with many coverage points, the computational load is large and the planning time is long, which not only increases the maintenance cycle but also increases the cleaning cost and affects the efficiency of laser cleaning. Summary of the Invention

[0005] The problem this invention aims to solve is how to improve the accuracy of laser cleaning path planning and improve the efficiency of laser cleaning in the composite path planning process of laser intelligent cleaning.

[0006] To address the above problems, this invention provides a composite path planning method for a laser intelligent cleaning process, comprising:

[0007] Obtain the binarized image corresponding to the workpiece to be cleaned, and divide the binarized image into regions to obtain multiple sub-regions;

[0008] Based on the set path heuristic method, path planning is performed on each of the sub-partitions to obtain the corresponding path planning data;

[0009] Based on the improved ant colony algorithm, path planning is performed on all the sub-partitions according to the path planning data of each sub-partition to obtain a global composite path plan.

[0010] Optionally, the path planning data includes the planned paths and path point sets of the corresponding sub-partitions; the improved ant colony algorithm, which performs path planning on all sub-partitions based on the path planning data of each sub-partition to obtain a global composite path plan, includes:

[0011] Based on the improved ant colony algorithm, path planning is performed on all the sub-partitions according to the set of path points to obtain a temporary path plan;

[0012] The global composite path plan is obtained based on the planned paths and temporary path plans of all the sub-partitions.

[0013] Optionally, the path point set includes combined nodes of each of the sub-partitions; the step of performing path planning on all the sub-partitions based on the improved ant colony algorithm, according to the path point set, to obtain a temporary path plan, includes:

[0014] Get the initial position;

[0015] Starting from the initial position, any ant in the initial ant colony visits different combined nodes, and the state transition probability is obtained based on the pheromone content on each combined node.

[0016] The temporary path plan is obtained based on the state transition probability and the corresponding extended passband constraint of the combined node;

[0017] The combined node is the set of the start and end points of the corresponding sub-partition.

[0018] Optionally, obtaining the temporary path plan based on the state transition probability and the corresponding extended passband constraint of the combined node includes:

[0019] The next path point for the ant transfer is determined based on the state transition probability. The next path point is the starting point or ending point of the combined node. The planned path direction of the corresponding sub-partition is determined based on the extended passband constraint of the corresponding combined node until the target position is reached.

[0020] Repeat the ant's path search process until the termination condition is met to obtain the temporary path plan.

[0021] Optionally, the extended passband constraint process includes:

[0022]

[0023]

[0024] Here, all the possible path points traversed by the ant during its visit are represented in sequence form, as follows: Let I be the set of positive integers, and R be the set of positive integers. i (1) R is the initial starting point set for the combined node i. i (2) S is the initial set termination point for the combined node j.j (1) S is the initial starting point set for the combined node j. j (2) The initial set termination point of the combined node j, R i For the combined node i, S j For the combined node j, This refers to the passband termination point of the combined node i after the extended passband constraint transformation. dir is the passband termination point of the combined node j after the extended passband constraint transformation. i and dir j These represent the directions of passband transmission.

[0025] Optionally, the state transition probability includes:

[0026]

[0027] Where, τ rs (t) represents the path l at time t. rs The amount of pheromones on, a μ Let μ be the set of possible path points for ant μ, q be the cumulative constant in the set of possible path points, and a be the set of possible path points. μ =C\t μ α and β are the pheromone heuristic factor and the expectation heuristic factor, respectively; η rs (t)=1 / d rs Let d be a heuristic function, where d rs η is the distance between adjacent path points. rq (t)=1 / d rq d is the accumulation constant of the heuristic function. rq The distance between r and q in the current accumulated beat. Let μ be the state transition probability of ant μ when visiting the path point.

[0028] Optionally, the step of accessing different combined nodes through any ant in the initial ant colony further includes:

[0029] When the ant visits the combined node, it updates the current combined node to the tabu list, and the pheromone update status of the current combined node is as follows:

[0030] τ rs (t+n)=(1-ρ)τ rs (t)+Δτ rs (t, t+n);

[0031]

[0032] Where, τrs (t+n) represents the path at time t+n. rs On the pheromone, τ rs (t) represents the path l at time t. rs The amount of pheromone on the surface, ρ is the pheromone volatility coefficient, ranging from [0,1], 1-ρ is the pheromone residue factor, and Δτ rs (t, t+n) represents the pheromone increment on the path (r, s) in the current loop, where Δτ is the initial value. rs (t, t+n) = 0, then Let denot be the pheromone increment left by the μ-th target ant on the access path (r,s) in this loop.

[0033] The advantages of the laser intelligent cleaning process composite path planning method described in this invention compared to existing technologies are as follows: By dividing the workpiece into regions and planning individual paths—that is, by dividing the binary image corresponding to the workpiece to be cleaned into regions and planning paths for each sub-region according to a set path heuristic—it is possible to design cleaning paths suitable for different regions (sub-regions) more precisely, thereby improving cleaning accuracy. Furthermore, by utilizing an improved ant colony algorithm for global path planning, it is possible to globally consider the path planning data of each sub-region, designing the optimal cleaning path for the entire workpiece, thus improving cleaning efficiency and quality. Therefore, this invention, by combining multiple path planning technologies and incorporating an improved ant colony algorithm in the global path access stage (global path planning), achieves global concatenation coverage of all sub-regions and can accurately and efficiently plan laser cleaning paths, thereby improving cleaning accuracy and efficiency.

[0034] To address the aforementioned technical problems, the present invention also provides a composite path planning system for a laser intelligent cleaning process, comprising:

[0035] The acquisition unit is used to acquire a binary image corresponding to the workpiece to be cleaned, and to divide the binary image into regions to obtain multiple sub-regions;

[0036] The processing unit is used to perform path planning for each of the sub-partitions according to the set path heuristic method to obtain the corresponding path planning data;

[0037] The processing unit is also used to perform path planning on all the sub-partitions based on the improved ant colony algorithm and the path planning data of each sub-partition, to obtain a global composite path plan.

[0038] The laser intelligent cleaning process composite path planning system and the laser intelligent cleaning process composite path planning method described in this invention have the same advantages over the prior art, and will not be repeated here. Attached Figure Description

[0039] Figure 1 This is a flowchart illustrating the composite path planning method for laser intelligent cleaning in an embodiment of the present invention.

[0040] Figure 2 This is a schematic diagram of the binarized image corresponding to the workpiece to be cleaned in an embodiment of the present invention;

[0041] Figure 3 This is a schematic diagram illustrating an ant traversing different sub-partitions during pathfinding, according to an embodiment of the present invention.

[0042] Figure 4 This is a schematic diagram illustrating the region division of a binarized image in an embodiment of the present invention;

[0043] Figure 5 This is a schematic diagram illustrating local path coverage of a sub-partition in an embodiment of the present invention;

[0044] Figure 6 This is a schematic diagram of global composite path planning in an embodiment of the present invention. Detailed Implementation

[0045] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0046] like Figure 1 As shown, in one embodiment, a composite path planning method for a laser intelligent cleaning process is provided, specifically including the following steps:

[0047] Step S1: Obtain the binarized image corresponding to the workpiece to be cleaned, and divide the binarized image into regions to obtain multiple sub-regions;

[0048] Specifically, the workpiece to be cleaned is typically a planar substrate of a certain size with rust or paint residue, made of materials such as steel, aluminum alloy, titanium alloy, or composite glass. The process of acquiring a binarized image of the workpiece can be achieved through digital imaging acquisition, image processing, and coordinate matching at different cleaning stages. For example, taking path planning for a rusted area on a steel alloy substrate as an example, the position of the workpiece to be cleaned is adjusted and calibrated to ensure that the workpiece surface can be fully distinguished. Graphical sampling and digital image processing are then performed on the workpiece surface to be cleaned to obtain a binarized image of the workpiece, such as... Figure 2 As shown, the white area is the target area to be zoned and planned, and the black area is the background area.

[0049] The binary image is divided into sections based on a set partitioning method to provide more accurate information for subsequent path planning. The cleaning stage includes cleaning the complete workpiece before cleaning and secondary laser cleaning of residual dirt after the initial cleaning. Only one image needs to be acquired for each cleaning stage.

[0050] The partitioning method can be selected from connected components, Boustrophedon (also known as the ox-plowing method), or clustering, etc. The specific partitioning method can be determined according to the actual situation. For example, when using the Boustrophedon cell decomposition method to perform region rule partitioning on a binarized image, the discrete unit of the image is first defined as an equivalent spot coverage unit (here, in the initial path planning stage, the matching process between pixel grid and spot size is omitted, and the focus is on the path planning process, where the size of a single pixel is equivalent to the size of the actual spot covering the sample surface, and path planning design is carried out with pixels as path units). The calculation results of partitioning the image (binarized image) using the Boustrophedon cell decomposition method are recorded in the form of a set of pixel coordinates, and finally, a two-dimensional space set of all sub-partitions (including the set of path points on the path) is obtained.

[0051] Among them, the Boustrophedon cell decomposition method is a partitioning method that uses the topological changes of scanned slices to determine the isolation of regions. It is represented as: using line segments as the slice function CS of the scan configuration space CS. λ λ is an adjustable parameter of the slice function, and the relationship between the slice function and the configuration space is expressed as CS = ∪ λ CS λ The change in λ is CS λ Implement the scanning process for CS. Then, let FS... λ A slice contained in free space is denoted as FS. λ =CS∩CS λ FS is determined by IN and OUT events. λ The changes in topology lead to the decomposition of old cells or the creation of new cells until the slicing function reaches the final scan boundary, at which point the entire configuration space is decomposed and no more cells are generated.

[0052] Step S2: Perform path planning on each sub-partition according to the set path heuristic method to obtain the corresponding path planning data;

[0053] Specifically, using a defined path heuristic, path planning is performed for each sub-region to generate cleaning path planning data suitable for each sub-region. This helps to design the optimal cleaning path for different regions based on actual conditions. Local path planning can be performed on each sub-region using a defined path heuristic (a heuristic for local path planning). This heuristic includes, but is not limited to, bow-shaped path coverage, loop-shaped path coverage, and inner spiral scan coverage. For example, taking the inner spiral path coverage method, path planning is performed on each sub-region separately, generating coverage paths for all sub-regions and obtaining the corresponding path index order within each sub-region (the index order is the path information of the sub-region's path planning data, i.e., the planned path and path point set of the sub-region; each path point is assigned a corresponding order, such as a label). The path information and the start and end points of the path are recorded. Furthermore, the inner spiral path heuristic method ensures the generation of more convex edge paths, maximizing the satisfaction of the convex angle path requirements for subsequent dynamic trajectory tracking of the path by the light spot. Repeat the above process one by one according to the partition order (the sub-partitions can be named, such as 1, 2, 3..., and the regions are divided one by one in order) until the path planning of all sub-partitions is completed and the corresponding path planning data is obtained.

[0054] Step S3: Based on the improved ant colony algorithm, perform path planning on all sub-partitions according to the path planning data of each sub-partition to obtain a global composite path plan;

[0055] Specifically, an improved ant colony algorithm is used to plan the coverage path for the global target area. This integrates path planning data generated from sub-regions, and sets appropriate parameters such as the initial number of ants, pheromone heuristic factor, expected heuristic factor, and pheromone evaporation rate. During the process, the target ants iteratively optimize the path until the optimization is complete, resulting in the optimal global coverage path, which represents the scanning coordinate sequence of the laser beam on the substrate. By comprehensively considering the path planning data from each sub-region through the improved ant colony algorithm, a comprehensive composite path plan is obtained, achieving more comprehensive and efficient laser cleaning path planning. The improved ant colony algorithm can effectively search for the optimal path across the entire workpiece.

[0056] The laser intelligent cleaning process composite path planning method described in this embodiment divides the workpiece into regions and plans individual paths. Specifically, it divides the binary image of the workpiece to be cleaned into regions and plans paths for each sub-region according to a set path heuristic. This allows for more precise design of cleaning paths suitable for different regions (sub-regions), improving cleaning accuracy. Furthermore, it utilizes an improved ant colony algorithm for global path planning, which globally considers the path planning data of each sub-region, designing the optimal cleaning path for the entire workpiece and improving cleaning efficiency and quality. Therefore, this invention combines multiple path planning technologies and incorporates an improved ant colony algorithm in the global path access stage (global path planning) to achieve global concatenation coverage of all sub-regions, enabling precise and efficient laser cleaning path planning, thus improving cleaning accuracy and efficiency.

[0057] In some embodiments, the path planning data includes the planned paths and path point sets of the corresponding sub-partitions; in step S3, the improved ant colony algorithm is used to perform path planning on all sub-partitions according to the path planning data of each sub-partition to obtain a global composite path plan, including:

[0058] Step S31: Based on the improved ant colony algorithm, perform path planning on all the sub-partitions according to the set of path points to obtain a temporary path plan;

[0059] Step S32: Obtain the global composite path plan based on the planned paths and temporary path plans of all the sub-partitions:

[0060] In some embodiments, the path point set includes combined nodes of each of the sub-partitions, and the combined node includes a start point and an end point. Step S31, based on the improved ant colony algorithm, performs path planning on all the sub-partitions according to the path point set to obtain a temporary path plan, including:

[0061] Step S311, obtain the initial position;

[0062] Step S312: Starting from the initial position, any ant in the initial ant colony visits different combined nodes, and the state transition probability is obtained based on the pheromone content of each combined node.

[0063] Step S313: Obtain the temporary path plan based on the state transition probability and the extended passband constraint of the corresponding combined node;

[0064] The combined node is the set of the start and end points of the corresponding sub-partition.

[0065] Specifically, an improved ant colony algorithm is used to perform global composite path planning that integrates local path heuristics (path planning data of sub-partitions). In this global composite path planning process, the improved ant colony algorithm addresses a global coverage problem with sub-partitions as special nodes (combined nodes). Each node is represented by either a marker point (i.e., a path point; in this case, the sub-partition has only one path coverage unit, where the path start point is also the end point) or two marker points (the path start point and end point within the sub-partition do not overlap). This node, as an independent unit that has achieved internal path coverage, possesses planar region characteristics and necessarily exhibits complete and reciprocal connectivity between its marker points (start and end points). The improved ant colony algorithm integrates these characteristics with the traditional ant colony algorithm, resulting in a composite ant colony algorithm (improved ant colony algorithm) based on locally reciprocal passband expansion. This makes the internal coverage path (planned path of the sub-partition) an extended transmission passband during the ant colony's access process. Furthermore, it utilizes the coexistence of multiple marker points within the combined node to increase the probability of ants randomly accessing the optimal pathfinding point.

[0066] Specifically, step-by-step path planning enables more accurate acquisition of temporary and global composite path plans, improving the precision and effectiveness of path planning. Path planning based on path point sets fully considers the specific needs of sub-regions, enhancing the flexibility and customizability of path planning. It also better utilizes the information from path point sets, fully considering the combination of nodes in different sub-regions. Through state transition probabilities and extended passband constraints, it more scientifically implements ant colony algorithm path search, improving the efficiency and quality of path planning. In other words, the improved ant colony algorithm, based on state transition probabilities and extended passband constraints, is more universal and flexible in its randomness. The combination of nodes increases the probability of the ant colony finding a better path distance by adding flexibility and randomness to path finding. Universality refers to a more compatible processing method for random scenarios, better adapting to the shape and characteristics of different workpieces, improving the adaptability and flexibility of path planning. Furthermore, the global-local composite path planning method saves computation time and reduces the overall path distance, improving the efficiency and quality of cleaning while achieving global composite path planning.

[0067] It should be noted that, for determining the starting and ending points in the composite node, they can first be set. The actual starting and ending points can be determined by using an improved ant colony algorithm and combined with extended passband constraints (that is, the starting point of the next composite node is determined by the state transition probability, and the internal movement direction of the sub-partition is determined by the extended passband constraints). The final global composite path planning can then be determined by using the planned path within the partition.

[0068] The improved ant colony algorithm is represented as follows:

[0069] Treating all sub-partitions as a series of combined nodes in the global plan (combined nodes include the start and end points of the planning path of the sub-partitions), let b i (t) represents the number of ants located in the i-th combination node at time t, n represents the number of combination nodes visited, and m represents the total number of ants in the ant colony.

[0070]

[0071] Equation (1) shows that in the improved ant colony algorithm, ants in the ant colony also complete the access between different combined nodes. First, it is determined that the actual path points of the ants are the marked points (start and end points or termination points) inside the combined node, rather than the combined node itself. It is represented as the path l between randomly marked points r and s at time t within the universal set C (the set of all combined nodes of sub-partitions) formed by any two combined nodes i and j. rs The set of residual pheromone amounts, where τ rs (t) represents the path l at time t. rs The pheromone content is determined by keeping the pheromone content of each path equal and constant at the initial time. Basic optimization is achieved through a directed graph g{C,L,Γ}, where L is the marker connection matrix.

[0072] In some embodiments, step S313, obtaining the temporary path plan based on the state transition probability and the corresponding extended passband constraint of the combined node, includes:

[0073] Step T1: Determine the next path point for the ant's transfer based on the state transition probability. The next path point is the starting point or ending point of the combined node. Determine the planned path direction of the corresponding sub-partition based on the extended passband constraint of the corresponding combined node until the target location is reached.

[0074] Step T2: Repeat the ant path search process until the termination condition is met, and obtain the temporary path plan.

[0075] Specifically, during the movement of μ (μ = 1, 2, ..., m) ants, the direction of movement is determined based on the amount of pheromone remaining on each path, and a taboo list t is used. μ To record the set of markers that the ants have traversed.

[0076] In some embodiments, the extended passband constraint process includes:

[0077]

[0078]

[0079] Here, all the possible path points traversed by the ant during its visit are represented in sequence form, as follows: Let I be the set of positive integers, and R be the set of positive integers. i (1) R is the initial starting point set for the combined node i. i (2) S is the initial set termination point for the combined node j. j (1) S is the initial starting point set for the combined node j. j (2) The initial set termination point of the combined node j, R i For the combined node i, S j For the combined node j, This refers to the passband termination point of the combined node i after the extended passband constraint transformation. dir is the passband termination point of the combined node j after the extended passband constraint transformation. i and dir j These represent the directions of passband transmission.

[0080] Specifically, R i and S j Let R be two sequences, namely the set of starting and ending points (combined nodes) of each sub-partition, and let I be the set of positive integers in mathematics (such as 1, 2, 3, ...). That is, the definition of the sequence includes the definition of all marker points (path points), i.e., R. i (1) R i (2) S j (1) S j (2) These represent the initial starting point (starting point) and ending point (ending point) of their respective sub-partitions.

[0081] in, Equations 1 and 2 here indicate that the ant's pathfinding is determined by the actual initial marker point in the sub-partition, and this serves as the constraint condition. After extending the passband constraint, the transformed termination marker point (starting and ending points) and the passband extension direction (passband transmission direction) are obtained.

[0082] In some embodiments, the state transition probability includes:

[0083]

[0084] Where, τ rs (t) represents the path l at time t. rs The amount of pheromones on, a μLet μ be the set of possible path points for ant μ, q be the cumulative constant in the set of possible path points, and a be the set of possible path points. μ =C\t μ α and β are the pheromone heuristic factor and the expectation heuristic factor, respectively; η rs (t)=1 / d rs Let d be a heuristic function, where d rs η is the distance between adjacent path points. rq (t)=1 / d rq d is the accumulation constant of the heuristic function. rq The distance between r and q in the current accumulated beat. Let μ be the state transition probability of ant μ when visiting the path point.

[0085] Specifically, taking two adjacent, already determined combination nodes i and j as an example, r and s are rewritten as a sequence of marked points with combination nodes i and j as the base, resulting in: Set the initial point (initial position) for global pathfinding. The starting access points for the two combined nodes are R. i S j Then, for the original state probability transition represented by equation (2), an extended passband constraint process is introduced, and the passband termination point in i,j is obtained through the transformations of equations (3) and (4). (The transformed termination point is obtained after extending the passband constraints) and the passband transmission direction dir i dir j A schematic diagram of the ant colony pathfinding process with integrated reciprocal passband expansion is shown below. Figure 4 As shown, R(S) indicates that the combined node is both the starting node of the ant colony in the current path traversal and the traversal node of the ant colony in the previous path traversal.

[0086] Figure 3 In this context, the main focus is on the handling of the target ant traversing different sub-partitions during pathfinding. The red and green dashed lines represent several traversal paths for the ant when traversing sub-partitions, including its starting point and visiting point. Taking the traversal between node a (sub-partition a) and node i (sub-partition i) as an example, if the ant's visiting path is the path represented by gray, that is, the ant's visiting point is the starting marker point in node i, then according to formula (3), a positive (represented by a gray composite arrow) extension passband is triggered, and the ant travels from R(S). i1 The marker point is transferred to R(S). i2Due to the reciprocity of passbands, the reverse propagation direction fails at the marker point. The transferred position is used as the starting marker point for the next search cycle in the next sub-partition ant pathfinding; if the ant's access path is the path indicated by black, that is, the ant's access point is the termination marker point in node i, then by formula (3), the reverse (indicated by black dashed and solid composite arrows) extended passband is triggered, and the ant moves from R(S). i2 The marker point is transferred to R(S). i1 Due to the inverse nature of passbands, the forward propagation direction fails for the marker point. Similarly, the transitioned position is used as the starting marker point for the next search cycle in the next sub-partition ant search. Figure 3 When a1 = a2, it means that the starting point and ending point of the sub-partition are the same. When a2 = a1 + 1, it means that the starting point and ending point of the sub-partition are different. Similarly, nodes i, j and n1 are similar to node a.

[0087] By using equations (3) and (4), and combining them... Figure 3 It can be seen that the optimal path point (combined node) of the ant colony between combined nodes is determined by the state transition function and the mutually inverse extended passband. Inside the combined node, the target ant's implementation through the mutually inverse extended passband is based on the control and reuse of the internal spiral heuristic algorithm. Due to its reversibility, the distance of the covered path in the sub-partition does not change during the control process, satisfying the global continuous transfer condition.

[0088] In some embodiments, the access between different combined nodes via any ant in the initial ant colony further includes:

[0089] When the ant visits the combined node, it updates the current combined node to the tabu list, and the pheromone update status of the current combined node is as follows:

[0090] τ rs (t+n)=(1-ρ)τ rs (t)+Δτ rs (t, t+n); (5)

[0091]

[0092] Where, τ rs (t+n) represents the path at time t+n. rs On the pheromone, τ rs (t) represents the path l at time t. rs The amount of pheromone on the surface, ρ is the pheromone volatility coefficient, ranging from [0,1], 1-ρ is the pheromone residue factor, and Δτ rs (t, t+n) represents the pheromone increment on the path (r, s) in the current loop, where Δτ is the initial value.rs (t, t+n) = 0, then Let denot be the pheromone increment left by the μ-th target ant on the access path (r,s) in this loop.

[0093] Specifically, after expanding the passband constraints, the target ant must update the tabu list t with all marked points (starting and ending points) within the combined node containing the visited path points. μ Meanwhile, considering the dynamic changes in ant colony access paths before and after being constrained by extended passbands, corresponding adjustments are also made during the pheromone update process.

[0094] Furthermore, due to the emphasis on the overall algorithmic nature of the Ant-Cycle model, variant models of Ant-Cycle can be used to update the overall pheromone levels. Regarding the changes in the original visitor locations of ants within the combined node before and after the extended passband constraint, the pheromone incremental update formula based on the Ant-Cycle variant is expressed as follows:

[0095]

[0096] In equation (7), Q is the pheromone intensity factor, and L μ Let be the path length that the μ-th ant traverses between the combined nodes in this loop.

[0097] After all iterations are completed, the optimal global path is obtained (the minimum distance can be selected as the optimal selection condition), and the local path points after passband constraint transformation are embedded into the global path to obtain a complete composite target coverage path (global composite path planning).

[0098] In one specific embodiment, the specific steps of the composite path planning method for the laser intelligent cleaning process are as follows:

[0099] Step A1, Digital Image Sampling and Processing: Adjust and calibrate the position of the workpiece to be cleaned to ensure that the workpiece surface can be completely distinguished. Perform graphic sampling and digital image processing on the workpiece surface to be cleaned to obtain the binarized image corresponding to the workpiece (e.g., ...). Figure 2 (as shown);

[0100] Step A2: Partitioning and path coverage of sub-partitions: The boustrophedon method is used to partition the binarized image, resulting in multiple sub-partitions. The number of partitions and their corresponding index information (named labels) are recorded. The initial scan point for each sub-partition is defined, typically at the edge point closest to the target centroid on the outermost edge of the sub-partition. An internal spiral heuristic is then used to generate the internal coverage path for each sub-partition, such as... Figure 4 and Figure 5 As shown, the display is differentiated by the edges of the sub-partitions. Figure 5 This means that the irregular target region in the binary image is divided into 95 regular sub-partitions, which means that the ant colony needs to traverse and visit 95 combined nodes, which are sequentially ordered by numerical sequence. Figure 5 The inner spiral scan method achieves path coverage from the starting point to the ending point of all original sub-partitions, and obtains the 186 marked points and their coordinates that the ant colony needs to visit while obtaining the local original coverage path and distance length.

[0101] Step A3 involves using an ant colony algorithm with fused reciprocal passband extension (an improved ant colony algorithm) to plan coverage paths for the global target region (sub-region). This integrates the combined node information generated from the sub-partitions, setting appropriate parameters such as the initial number of ants, pheromone heuristic factor, expected heuristic factor, and pheromone evaporation rate. During the process, the target ants iteratively optimize the path periodically. Simultaneously, the path coverage direction within the corresponding sub-partition is defined based on the order of marker points within the same combined node, until the ant iteration optimization is complete, resulting in the optimal global coverage path, which is the scanning coordinate order of the laser beam on the action substrate. The ant colony algorithm settings and path distance calculation results are shown in Table 1. To visually demonstrate the global coverage path, all paths are depicted based on the partition diagram. The global coverage diagram is shown below. Figure 6 As shown.

[0102] Table 1. Ant colony algorithm settings and path distance calculation results

[0103] Indicator parameters Quantity Ant count 15 Pheromones as heuristics 0.9 Expected heuristic factor 3 Pheromones evaporation rate 0.1 Total distance of spiral coverage within sub-partition 31954 Ant colony global planning distance 1520 Total distance 33474

[0104] This invention utilizes the Boustrophedon unit decomposition method to divide the cleaned target into sub-regions after converting it into a binary image, and extracts the sub-region sequence. Starting points for each sub-region are defined, and heuristic path filling based on the inner spiral method is performed. Combined path points supporting inter-bandgap transmission within the sub-region are obtained based on the heuristic paths. In the global path access stage, the ant colony algorithm is improved to establish a global ant colony path planning constrained by local reciprocal passbands, thereby achieving global concatenated coverage of all sub-regions. This invention enables precise and efficient path optimization for random target regions, significantly reducing scanning path distance, and is intuitive and convenient to operate, saving manpower and time costs.

[0105] The laser intelligent cleaning process composite path planning method described in this embodiment divides the workpiece into regions and plans individual paths. Specifically, it divides the binary image of the workpiece to be cleaned into regions and plans paths for each sub-region according to a set path heuristic. This allows for more precise design of cleaning paths suitable for different regions (sub-regions), improving cleaning accuracy. Furthermore, it utilizes an improved ant colony algorithm for global path planning, which globally considers the path planning data of each sub-region, designing the optimal cleaning path for the entire workpiece and improving cleaning efficiency and quality. Therefore, this invention combines multiple path planning technologies and incorporates an improved ant colony algorithm in the global path access stage (global path planning) to achieve global concatenation coverage of all sub-regions, enabling precise and efficient laser cleaning path planning, thus improving cleaning accuracy and efficiency.

[0106] Corresponding to the aforementioned composite path planning method for laser intelligent cleaning processes, this invention also provides a composite path planning system for laser intelligent cleaning processes. The composite path planning system for laser intelligent cleaning processes includes:

[0107] The acquisition unit is used to acquire a binary image corresponding to the workpiece to be cleaned, and to divide the binary image into regions to obtain multiple sub-regions;

[0108] The processing unit is used to perform path planning for each of the sub-partitions according to the set path heuristic method to obtain the corresponding path planning data;

[0109] The processing unit is also used to perform path planning on all the sub-partitions based on the improved ant colony algorithm and the path planning data of each sub-partition, to obtain a global composite path plan.

[0110] It should be noted that, in this document, terms such as "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0111] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

[0112] While the present invention has been disclosed above, its scope of protection is not limited thereto. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention, and all such changes and modifications will fall within the scope of protection of the present invention.

Claims

1. A composite path planning method for a laser intelligent cleaning process, characterized in that, include: Obtain the binarized image corresponding to the workpiece to be cleaned, and divide the binarized image into regions to obtain multiple sub-regions; Based on the set path heuristic method, path planning is performed on each of the sub-partitions to obtain the corresponding path planning data; Based on the improved ant colony algorithm, path planning is performed on all the sub-partitions according to the path planning data of each sub-partition to obtain a global composite path plan; The path planning data includes the planned paths and path point sets for the corresponding sub-partitions; the improved ant colony algorithm performs path planning on all sub-partitions based on the path planning data of each sub-partition to obtain a global composite path plan, including: Based on the improved ant colony algorithm, path planning is performed on all the sub-partitions according to the set of path points to obtain a temporary path plan; The global composite path plan is obtained based on the planned paths and temporary path plans of all the sub-partitions; The path point set includes combined nodes of each of the sub-partitions; the improved ant colony algorithm, based on the path point set, performs path planning on all the sub-partitions to obtain a temporary path plan, including: Get the initial position; Starting from the initial position, any ant in the initial ant colony visits different combined nodes, and the state transition probability is obtained based on the pheromone content on each combined node. The temporary path plan is obtained based on the state transition probability and the corresponding extended passband constraint of the combined node; Wherein, the combined node is the set of the start and end points of the corresponding sub-partition; The step of obtaining the temporary path planning based on the state transition probability and the corresponding extended passband constraint of the combined node includes: The next path point for the ant transfer is determined based on the state transition probability. The next path point is the starting point or ending point of the combined node. The planned path direction of the corresponding sub-partition is determined based on the extended passband constraint of the corresponding combined node until the target position is reached. Repeat the ant's path search process until the termination condition is met to obtain the temporary path plan.

2. The composite path planning method for laser intelligent cleaning process according to claim 1, characterized in that, The extended passband constraint process includes: ； ； Wherein, all the selectable path points traversed by the ant during its visit are represented in sequence form, as follows: , Let I be the set of positive integers, and R be the set of positive integers. i (1) R is the initial starting point set for the combined node i. i (2) S is the initial set termination point for the combined node i. j (1) S is the initial starting point set for the combined node j. j (2) The initial set termination point of the combined node j, R i For the combined node i, S j For the combined node j, To extend the passband constraint transformation, the passband termination point of the combined node i is... To extend the passband constraint transformation, the passband termination point of the combined node j is... and These represent the directions of passband transmission.

3. The composite path planning method for laser intelligent cleaning process according to claim 1, characterized in that, The state transition probabilities include: ； in, for Time Path Pheromones on the surface For ants The next option is the set of path points. This is an accumulated constant within the set of selectable path points. ; and These are the pheromone heuristic factor and the expectation heuristic factor, respectively. Here is the heuristic function, where The distance between adjacent path points. This is the accumulation constant of the heuristic function. For the path points in the current accumulated beat and The distance between them For ants The state transition probability when visiting the path point. Let be the pheromone concentration on the path from the current point r to the optional point q at time t, and let s be the candidate point currently being considered as the next path point.

4. The composite path planning method for laser intelligent cleaning process according to claim 1, characterized in that, The method of accessing different combined nodes through any ant in the initial ant colony also includes: When the ant visits the combined node, it updates the current combined node to the tabu list, and the pheromone update status of the current combined node is as follows: ； ； in, for Time Path On the pheromones, for Time Path Pheromones on the surface The pheromone evaporation coefficient varies from [0,1]. The residual factor of the pheromone, This indicates the path visited in the current loop. The pheromone increment on the surface, at the initial time ,but Represented as the first Only the target ant remains on the visiting path in this loop. The pheromone increment is given by m, where m represents the total number of ants in the ant colony.

5. A composite path planning system for a laser intelligent cleaning process, characterized in that, The laser intelligent cleaning process composite path planning method as described in any one of claims 1 to 4, wherein the planning system comprises: The acquisition unit is used to acquire a binary image corresponding to the workpiece to be cleaned, and to divide the binary image into regions to obtain multiple sub-regions; The processing unit is used to perform path planning for each of the sub-partitions according to the set path heuristic method to obtain the corresponding path planning data; The processing unit is also used to perform path planning on all the sub-partitions based on the improved ant colony algorithm and the path planning data of each sub-partition, to obtain a global composite path plan.

Images