A plywood automatic repairing method based on heuristic algorithm

Abstract

The present application relates to the technical field of plywood defect repair, and specifically relates to a plywood automatic repair method based on a heuristic algorithm, which comprises a conveying belt for conveying plywood, a feeding area and a stacking area respectively arranged along the front and rear ends of the conveying belt, a visual detection system module and at least one mechanical hand module sequentially arranged along the feeding direction of the conveying belt, the conveying belt, the visual detection system module and the mechanical hand module being respectively connected with and controlled by a central control system, and an automatic repair path planning module being arranged in the central control system. The defect repair method provided in the present application can realize rapid automatic defect repair of large-scale plywood in a moving state, can repair defects covering almost all possible defect types on the plywood, and finally realizes automatic repair of defects of the plywood, so as to improve the use performance of the plywood. In addition, the method can greatly save resources such as manpower, time and raw materials.

Description

Technical Field

[0001] This invention relates to the field of plywood defect repair technology, and in particular to an automated plywood repair method based on heuristic algorithms. Background Technology

[0002] Plywood is the most consumed type of board in the timber industry. Its production involves deep processing of wood, using a rotary cutter to slice veneers, which are then glued, spliced, adhesive-coated, assembled, hot-pressed, patched, trimmed, and sanded – a complex process to create a engineered wood product. During manufacturing, defects such as cracks and holes are present in the boards, and gaps or holes are easily created during splicing. These cracks and holes can be filled with putty to improve the performance and water resistance of the plywood.

[0003] The process of filling defects in plywood with putty is usually done manually. Workers use scrapers to apply putty into the defects by pressing it in, but manual work is inefficient and scraper application often consumes too much material, resulting in significant additional waste.

[0004] Research on automated repair of plywood defects is limited both domestically and internationally. Existing automated putty spraying equipment typically relies on large-diameter pumps to force putty into the defects under high pressure. This method usually only addresses defects on the sides of the plywood, and the filling process requires manual control. Furthermore, this method results in significant waste due to high wear and tear.

[0005] Therefore, a new technical solution is urgently needed to solve the aforementioned technical problems. Summary of the Invention

[0006] The purpose of this invention is to overcome the problems of the prior art and provide an automated plywood repair method based on heuristic algorithms to solve the technical problems of automated repair of plywood defects, reducing manpower and material consumption and saving time.

[0007] The above objectives are achieved through the following technical solutions: An automated plywood repair method based on heuristic algorithms includes a conveyor belt for transporting plywood, with a feeding area and a stacking area respectively arranged at the front and rear ends of the conveyor belt. A vision inspection system module and at least one robotic arm module are sequentially arranged along the feeding direction of the conveyor belt. The conveyor belt, the vision inspection system module, and the robotic arm module are connected to and controlled by a central control system, which includes an automatic repair path planning module. The steps are as follows: In step (1), the plywood enters the conveyor belt from the loading area. When the plywood is conveyed to the detection area of ​​the vision inspection system module, the vision inspection system module acquires images of the plywood in the moving state and uses a visual image algorithm to identify the number of defects, defect coordinates and defect size on the plywood, and generates defect data to be transmitted to the automatic repair path planning module. In step (2), the automatic repair path planning module receives the defect data, classifies it according to the defect shape (including point type, line type, and rectangle type), and calculates the deviation value caused by the movement of the conveyor belt. It then uses an automatic path planning algorithm based on heuristics to calculate the defect repair path of the corresponding robot arm in each robot arm module when the plywood passes through each robot arm module, so that the central control system can control the path planning of each robot arm. Step (3) When the plywood is conveyed to the working area corresponding to the robot module, the robot performs defect repair operation on the plywood in the moving state according to the defect repair path allocated by the central control system. Step (4) After the plywood is repaired, it is directly conveyed to the stacking area for stacking, and the unloading is completed.

[0008] Furthermore, step (2) specifically includes: In step (2-1), the automatic repair path planning module receives the defect data, calculates the repair time required to repair a single defect, and determines the number of robotic arm modules that need to participate in defect repair. Step (2-2) uses an automatic path planning algorithm based on heuristics to plan the path for the defect data on a single sheet of plywood and calculates the repair time required for all defects on a single sheet of plywood. In step (2-3), the central control system sends the path planning information from step (2-2) to the designated robotic arm module, and the robotic arm in the robotic arm module performs repair work based on the received path planning information. In step (2-4), the central control system stores the path planning information described in step (2-2), with each plywood sheet serving as a storage unit.

[0009] Further, in step (2-2), the automatic path planning algorithm based on heuristics is used to plan the path for the defect data on a single sheet of plywood. Specifically, this involves applying nonlinear constraints with the number of repairable defects as the objective function; the constraints include: Constraint 1: Path coordinate constraint, the x-axis corresponds to the length of the plywood, and the y-axis corresponds to the width of the plywood; Constraint 2: Time constraint, the time required for the robotic arm to repair to the current coordinate point; Constraint 3: Robotic arm performance constraints: Robotic arm speed < 1600 mm / s; Constraint 4: Mechanical constraint. The turning speed of the robot arm is determined by the turning angle and the length of the turning arc segment, and is a value less than 1600.

[0010] Furthermore, the time it takes for the robotic arm to repair the defect to the current coordinate point includes the lower bound of the repairable time range of the defect and the upper bound of the repairable time range of the defect.

[0011] Furthermore, in step (2-3), the robot arm does not need to lift its Z-axis after each defect is repaired during the repair process of each plywood sheet.

[0012] Furthermore, after the robot arm has finished repairing all defects on each plywood sheet as described in steps (2-3), the Z-axis needs to be lifted and the robot arm needs to switch to repairing the next plywood sheet.

[0013] Furthermore, in steps (2-3), after each plywood is repaired, the robotic arm plans the repair path for the next plywood based on the end point, without needing to plan based on the origin point.

[0014] Furthermore, there are three robotic arm modules, including robotic arm module 1, robotic arm module 2 and robotic arm module 3, and each is controlled by the central control system. Beneficial effects

[0015] This invention provides an automated plywood repair method based on a heuristic algorithm, and a corresponding intelligent system. This system uses a robotic arm to automatically repair plywood, reducing manpower consumption, saving time, and minimizing material waste. Furthermore, this method enables rapid and automated defect filling in large-scale plywood movement, covering almost all possible defect types found in plywood. Ultimately, it achieves automated defect repair of plywood, improving the performance of the boards and significantly saving manpower, time, and raw materials. Attached Figure Description

[0016] Figure 1 This invention relates to a defect repair system used in an automated plywood repair method based on a heuristic algorithm. Figure 2 This is a flowchart of the defect repair path planning in the automated plywood repair method based on heuristic algorithms described in this invention. Detailed Implementation

[0017] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments. The described embodiments are merely some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] like Figure 1 As shown, this solution provides an automated plywood repair method based on a heuristic algorithm, and prioritizes the design of a system for implementing this method, including a conveyor belt for transporting plywood, with a feeding area and a stacking area set at the front and rear ends of the conveyor belt, respectively. A vision inspection system module and at least one robotic arm module are arranged sequentially along the feeding direction of the conveyor belt. The conveyor belt, the vision inspection system module, and the robotic arm module are respectively connected to and controlled by a central control system, which includes an automatic repair path planning module.

[0019] Specifically, the vision inspection system module is positioned directly above the conveyor belt in the feeding direction. It is used to take pictures of the moving plywood to obtain images, which facilitates the extraction of defect parameters. That is, a vision algorithm is used to call an AI algorithm to identify defects and obtain defect coordinates. The images are divided according to color blocks, with light-colored parts representing the surface of the wood board and dark-colored parts representing defects that need repair. Based on the dark-colored areas, defects are categorized as points, blocks, lines, etc., and the coordinates are calculated based on their relative positions on the board.

[0020] The robotic arm module includes an active area corresponding to the conveyor belt, and its robotic arm can perform three-dimensional movement along the X, Y, and Z axes in order to receive instructions from the central control system to perform defect repair operations.

[0021] The automatic repair path planning module is a software program installed in the central control system, used to process the plywood parameters obtained by the vision inspection system module for path planning.

[0022] The system's workflow is as follows: Plywood is typically rectangular, with a length and width of less than 2.6 meters, and a thickness ranging from 5 mm to 25 mm. It is loaded by moving the boards onto a conveyor belt manually or by an automatic feeding machine (a type of lifting machine that uses cylinders for lifting).

[0023] The conveyor belt runs at a speed not exceeding 15 m / min; First, the plywood will be inspected by a vision inspection system module, which will automatically identify the location and type of defects on the plywood, as well as the defect attributes (such as length, width, and whether it is located on the edge of the board). The plywood is then conveyed through the range of a robotic arm (a device that can move within a three-dimensional cuboid space and is connected to a discharge pump). Through a plywood automatic repair path planning system based on heuristic algorithms, the optimal repair path of the robotic arm is dynamically calculated. The robotic arm drives the pump to evenly fill the defective locations of the plywood with putty, thus achieving automated repair of the plywood. The repaired plywood will be automatically unloaded from the conveyor belt and placed in the stacking area at the end of the conveyor belt, thus completing the unloading process.

[0024] The defects of the plywood described in this embodiment mainly include: 1. Indentations (dents): These generally refer to dents on the surface of the substrate being inspected caused by external pressure, with a depth not exceeding the thickness of the veneer. They may also be sheet-like dents caused by the inherent properties of the wood during rotary cutting, typically appearing in patches with irregular shapes, a width of less than 5mm, and a depth of less than 1mm. 2. Holes: Generally refers to a defect in which the surface details of the substrate being inspected are completely or partially detached. The major and minor diameters are generally within 30mm, and the ratio of the major to minor diameters is generally no greater than 3. 3. Cracks: These generally refer to fissures formed on the surface of the repaired substrate due to external forces or changes in temperature and humidity, causing the wood fibers to separate. The length is typically less than 600mm. The width is wider at the edges and narrower in the middle. Most cracks occur at the edges of the board, with a few in the middle, usually less than 100mm in length and less than 1mm in width. 4. Core separation: Generally refers to the gap formed when two veneers are not placed tightly during the assembly process. The gap is the length of the entire veneer and the width is generally within 5mm, with a few extreme cases reaching 10mm. 5. Overlapping core: refers to the defect caused by two veneers overlapping during the blank assembly process; 6. Diagonal grain: Irregular grain patterns produced during wood processing, generally appearing in patches, with a depth not exceeding 0.6mm, and the spacing between diagonal grains is generally between 2-5mm; 7. Missing edges and corners: This generally refers to the loss of edges and corners of a single board during assembly, cold pressing, hot pressing, or transfer; it is a defect that is connected to the edge and is missing in whole or in part, and the width is generally more than 3mm.

[0025] By classifying the above defect types into point types, line types, and rectangle types according to defect size, the robot arm is controlled to fill as many defects as possible along a certain trajectory. The filling schemes are divided into point filling, line filling, and reciprocating filling, each corresponding to one of the three defect types.

[0026] This solution can fill defects covering almost all types of defects that may appear in plywood, ultimately achieving automated repair of plywood defects and greatly saving manpower, time, raw materials and other resources.

[0027] In addition to repairing defects, a human-machine interface system can control functions such as conveyor belt speed, material flow rate at the outlet, robotic arm repair speed, and robotic arm pressure. This system can meet the needs of plywood repair in various situations and achieve more intelligent automated control.

[0028] This solution provides an automated plywood repair method based on a heuristic algorithm, with the following steps: In step (1), the plywood enters the conveyor belt from the loading area. When the plywood is conveyed to the detection area of ​​the vision inspection system module, the vision inspection system module acquires images of the plywood in the moving state and uses a visual image algorithm to identify the number of defects, defect coordinates and defect size on the plywood, and generates defect data to be transmitted to the automatic repair path planning module. In step (2), the automatic repair path planning module receives the defect data, classifies it according to the defect shape (including point type, line type, and rectangle type), and calculates the deviation value caused by the movement of the conveyor belt. It then uses an automatic path planning algorithm based on heuristics to calculate the defect repair path of the corresponding robot arm in each robot arm module when the plywood passes through each robot arm module, so that the central control system can control the path planning of each robot arm. Step (3) When the plywood is conveyed to the working area corresponding to the robot module, the robot performs defect repair operation on the plywood in the moving state according to the defect repair path allocated by the central control system. Step (4) After the plywood is repaired, it is directly conveyed to the stacking area for stacking, and the unloading is completed.

[0029] Furthermore, step (2) specifically includes: In step (2-1), the automatic repair path planning module receives the defect data, calculates the repair time required to repair a single defect, and determines the number of robotic arm modules that need to participate in defect repair; (in principle, the repair time required for a defect should not exceed the time required for the vision inspection system module to detect the size). Step (2-2) uses an automatic path planning algorithm based on heuristics to plan the path for the defect data on a single plywood sheet and calculates the repair time required for all defects on a single plywood sheet; the repair time for each board defect should be less than the defect detection time. In step (2-3), the central control system sends the path planning information from step (2-2) to the designated robotic arm module, and the robotic arm in the robotic arm module performs repair work based on the received path planning information. In step (2-4), the central control system stores the path planning information described in step (2-2), with each plywood sheet serving as a storage unit; this facilitates future path planning self-adaptation and self-learning, with a storage time of no less than 30 days and the storage format being IPG compression.

[0030] Step (2-2) involves using an automatic path planning algorithm based on heuristics to plan the path for the defect data on a single sheet of plywood. Specifically, this involves applying nonlinear constraints with the number of repairable defects as the objective function; the constraints include: Constraint 1: Path coordinate constraint, the x-axis corresponds to the length of the plywood, and the y-axis corresponds to the width of the plywood; Constraint 2: Time constraint, the time for the robot to repair to the current coordinate point, including the lower bound of the repairable time range of the defect point and the upper bound of the repairable time range of the defect point; Constraint 3: Robotic arm performance constraints: Robotic arm speed < 1600 mm / s; Constraint 4: Mechanical constraint. The turning speed of the robot arm is determined by the turning angle and the length of the turning arc segment, and is a value less than 1600.

[0031] The path planning requirements for this solution are as follows: 1) The plywood automatic repair path planning system uses the defect coordinates and size information of each plywood continuously detected by the vision inspection system module to plan the optimal path, so as to ensure that the robot arm repairs the plywood with the least time, the optimal path, and the best surface quality. 2) Path planning and calculation time should not exceed 0.5 seconds; 3) The communication protocol between the path planning system (automatic path repair module) and the automatic identification system is IP / TCP; 4) During the repair process of each board, after each defect is repaired, the Z-axis (i.e., the coordinate axis for the robot to move up and down) does not need to be raised; 5) After the robot arm finishes repairing each board, the Z-axis needs to be lifted to switch to the next board for repair; 6) After each board is repaired, the robotic arm plans the repair path for the next board using the end point as the reference point, without needing to use the origin point as the reference point.

[0032] As a specific embodiment of this solution, there is one robotic arm module, including robotic arm module 1, robotic arm module 2, and robotic arm module 3, each controlled by the central control system. The defect repair process for plywood is as follows: Figure 2 As shown.

[0033] Furthermore, the extraction, analysis, and planning of repair paths for plywood defects is a very complex process, as follows: The conveyor belt operates at a certain speed (typically 160-250 mm / s), and it does not stop during the repair process, while the robotic arm's range of motion is fixed. Therefore, during the repair process, the robotic arm may be forced to stop at its boundary (hereinafter referred to as "limit collision"). The limit collision problem depends on various factors, including the time required for the robotic arm to repair the defect, the conveyor belt speed, the length of the plank, the initial position of the robotic arm, the robotic arm's movement speed, and the robotic arm's filling speed. How to plan a reasonable path that meets several constraints, including: minimizing the robot's action time; ensuring the robot's range of motion does not exceed its limits; maintaining the robot within the wooden board area during the repair process and for a short period before and after repair (because filling defects requires pumping putty into the defects, the robot exerts significant downward pressure on the board; if it encounters a moving conveyor belt, it will severely damage the robot); and minimizing material waste. Furthermore, the robot's movement is not uniform; it accelerates and decelerates in straight lines, and its speed varies depending on the angle at corners. Therefore, the repair time cannot be simply calculated by dividing the path by the speed. Visual data is not always accurate. The coordinates of defects obtained through image algorithms often contain certain errors. These errors need to be reduced through complex formulas so as not to affect the repair effect of the robotic arm. The trajectory needs to be calculated as quickly as possible. The trajectory calculation time is strictly limited and must be completed within the time it takes for the wooden board to pass through the vision module and enter the processing range of the robotic arm. The vision module also needs a certain amount of time for recognition. Therefore, the entire system has very stringent time requirements, and the calculation time should be guaranteed to be within 0.5 seconds. Thus, there are very high requirements for the time complexity of the algorithm.

[0034] This invention innovatively uses a visual image algorithm to solve the problem of plywood defect identification, and uses the algorithm to convert image data into coordinates, which are then provided to the subsequent path planning system.

[0035] The path planning algorithm of this invention is based on a genetic algorithm, which improves its robustness and stability by reducing its random factor; and obtains the final algorithm evaluation method through a posterior fitness function, and then obtains a better trajectory plan based on the function; and improves its solution speed in large samples by optimizing the algorithm logic, so that it can give a better trajectory within 0.5 seconds, which meets the actual needs of the factory.

[0036] The path planning in this invention is achieved through a heuristic algorithm, specifically a modified Optimal Genetic Algorithm (OGA). A Genetic Algorithm (GA) is a computational model that simulates the biological evolutionary process, including natural selection and genetic mechanisms, and is a method for searching for optimal solutions by simulating natural evolution. This algorithm uses mathematical methods and computer simulations to transform the problem-solving process into processes similar to crossover and mutation of chromosomes and genes in biological evolution. Its main advantages include: 1) If handled properly, the global optimal solution can be obtained in the optimization problem; 2) Parallel computation is relatively easy to implement; 3) Genetic algorithms are result-oriented, focusing primarily on two aspects: encoding and decoding, and the fitness function. GA requires less attention to the process and has a very wide range of applications. It can solve linear, nonlinear, and discontinuous problems, and even some problems where the process is not clearly understood.

[0037] However, it also has some drawbacks. For example, it relies excessively on the fitness function, which essentially determines the outcome. This function needs to be provided by the user and lacks a standard, requiring development based on the specific implementation problem. Furthermore, improper population size selection or settings can lead to premature convergence, missing the global optimum and only reaching a local optimum. Because genetic algorithms don't focus heavily on the process, slight differences in parameters during computation can yield different results. These parameters include population size, crossover frequency, and mutation frequency. Additionally, when chromosomes are too long (i.e., too many genes), the number of recombinant segments increases dramatically, resulting in a very large population size and a sharp increase in computation time.

[0038] When solving complex combinatorial optimization problems, GA (Gaussian Algorithm) can typically achieve better optimization results faster than some conventional optimization algorithms. Traditional GA algorithms are usually based on several operators and create random factors through large-scale random mutation and crossover processes to obtain a relatively good or optimal solution. However, in solving this problem, the time required to repair each defect point needs to be calculated by planning the entire trajectory (because the robot's running speed is different for different trajectories), making it difficult to predict the time consumption of a trajectory. Furthermore, setting the fitness function under such conditions is a complex problem. In addition, when facing hundreds of defect points and hundreds of random gene combinations, under factory computer conditions, GA often takes more than 15 minutes to obtain an optimal solution, which far exceeds the 0.5-second threshold. Therefore, this invention improves upon the traditional GA algorithm by reducing the randomness factor and enhancing its robustness to meet the stability requirements of industrial scenarios. This allows the algorithm to plan a unique and relatively stable trajectory for each plywood plank. Furthermore, the solution speed is optimized specifically for this problem by employing an embedded roulette wheel strategy for gene combination transformation, satisfying the randomness requirement of the solution and significantly reducing computation time. Adaptation to posterior fitness has also been implemented, resulting in an OGA method with extremely fast computation time (less than 0.5 seconds for solving trajectories with over 100 points on a 10th-generation i5 processor) and stable solution (including consistently providing optimal solutions and yielding identical results for the same plywood plank, facilitating later verification). This method ensures that a robotic arm's trajectory is planned within 0.5 seconds of visual detection of defects on the plywood surface, guaranteeing that as many defective points as possible are repaired as possible when the plywood passes through the robotic arm's range.

[0039] The results obtained from the above algorithm are fed into the robot control program, and control settings including pump switch, speed control, limit protection, and pump material flow are added to form a complete plywood automated repair system that can detect plywood defects in real time and quickly perform automated repairs, greatly saving manpower, time costs, and material losses.

[0040] The above description is merely illustrative of the embodiments of the present invention and is not intended to limit the present invention. For those skilled in the art, any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An automated plywood repair method based on heuristic algorithms, characterized in that, The system includes a conveyor belt for transporting plywood, with a feeding area and a stacking area located at the front and rear ends of the conveyor belt, respectively. A vision inspection system module and at least one robotic arm module are sequentially arranged along the feeding direction of the conveyor belt. The conveyor belt, the vision inspection system module, and the robotic arm module are connected to and controlled by a central control system, which includes an automatic repair path planning module. The steps are as follows: In step (1), the plywood enters the conveyor belt from the loading area. When the plywood is conveyed to the detection area of ​​the vision inspection system module, the vision inspection system module acquires images of the plywood in the moving state and uses a visual image algorithm to identify the number of defects, defect coordinates and defect size on the plywood, and generates defect data to be transmitted to the automatic repair path planning module. In step (2), the automatic repair path planning module receives the defect data, classifies it according to the defect shape (including point type, line type, and rectangle type), and calculates the deviation value caused by the movement of the conveyor belt. It then uses an automatic path planning algorithm based on heuristics to calculate the defect repair path of the corresponding robot arm in each robot arm module when the plywood passes through each robot arm module, so that the central control system can control the path planning of each robot arm. Step (3) When the plywood is conveyed to the working area corresponding to the robot module, the robot performs defect repair operation on the plywood in the moving state according to the defect repair path allocated by the central control system. Step (4) After the plywood is repaired, it is directly conveyed to the stacking area for stacking, and the unloading is completed; Step (2) specifically includes: In step (2-1), the automatic repair path planning module receives the defect data, calculates the repair time required to repair a single defect, and determines the number of robotic arm modules that need to participate in defect repair. Step (2-2) uses an automatic path planning algorithm based on heuristics to plan the path for the defect data on a single sheet of plywood and calculates the repair time required for all defects on a single sheet of plywood. In step (2-3), the central control system sends the path planning information from step (2-2) to the designated robotic arm module, and the robotic arm in the robotic arm module performs repair work based on the received path planning information. In step (2-4), the central control system stores the path planning information described in step (2-2), with each plywood sheet serving as a storage unit.

2. The automated plywood repair method based on a heuristic algorithm according to claim 1, characterized in that, Step (2-2) involves using an automatic path planning algorithm based on heuristics to plan the path for the defect data on a single sheet of plywood. Specifically, this involves applying nonlinear constraints with the number of repairable defects as the objective function; the constraints include: Constraint 1: Path coordinate constraint, the x-axis corresponds to the length of the plywood, and the y-axis corresponds to the width of the plywood; Constraint 2: Time constraint, the time required for the robotic arm to repair to the current coordinate point; Constraint 3: Robotic arm performance constraints: Robotic arm speed < 1600 mm / s; Constraint 4: Mechanical constraint. The turning speed of the robot arm is determined by the turning angle and the length of the turning arc segment, and is a value less than 1600.

3. The automated plywood repair method based on a heuristic algorithm according to claim 2, characterized in that, The time it takes for the robotic arm to repair the defect to the current coordinate point includes the lower bound of the time range within which the defect can be repaired, and the upper bound of the time range within which the defect can be repaired.

4. The automated plywood repair method based on a heuristic algorithm according to claim 1, characterized in that, During the repair process of each plywood sheet, as described in steps (2-3), the Z-axis does not need to be lifted after each defect is repaired.

5. The automated plywood repair method based on a heuristic algorithm according to claim 1, characterized in that, After the robot arm has finished repairing all defects on each plywood sheet as described in steps (2-3), the Z-axis needs to be lifted and the robot arm needs to switch to repairing the next plywood sheet.

6. The automated plywood repair method based on a heuristic algorithm according to claim 1, characterized in that, In steps (2-3), after each plywood is repaired, the robotic arm plans the repair path for the next plywood based on the end point, without needing to plan based on the origin point.

7. The automated plywood repair method based on a heuristic algorithm according to claim 1, characterized in that, There are three robotic arm modules, each controlled by the central control system.

Images