Casting cleaning robot autonomous decision-making auxiliary system based on artificial intelligence

Abstract

The invention relates to the technical field of casting cleaning, and particularly discloses a casting cleaning robot autonomous decision-making auxiliary system based on artificial intelligence, and the system comprises a parameter obtaining module which obtains sub-block information, calculates the regional flatness, divides a sub-block into a priority processing block and a to-be-processed block, and obtains the priority processing block and the to-be-processed block; a to-be-cleaned block is further determined through a distance calculation module, a path planning module constructs an AI cleaning path model by using a clustering algorithm and obtains a cleaning path, and a cleaning decision module calculates a cleaning time ratio and a path cleaning ratio and obtains a decision path coefficient so as to determine an optimal cleaning path. And finally, the casting cleaning robot is guided to clean the casting surface according to the optimal path, the cleaning efficiency and quality are effectively improved, and the decision-making ability of the casting cleaning robot is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of casting cleaning, and in particular to an autonomous decision-making auxiliary system of a casting cleaning robot based on artificial intelligence. Background Art

[0002] In the foundry industry, casting cleaning is a key link in the production process. The traditional casting cleaning method mainly relies on manual operation, but this method has many disadvantages.

[0003] In the prior art, large castings are divided into several units for separate grinding, and the castings or robots are frequently moved to switch grinding units. This is not only cumbersome and inefficient, but also unable to quickly analyze the castings and automatically plan efficient grinding paths. In terms of grinding quality, the prior art relies on manual drag-and-drop teaching or offline programming, which can easily lead to unstable quality due to differences in manual operations and inconsistencies between the model and the actual situation, and does not use intelligent control algorithms to monitor the flatness of the casting surface in real time. For operators, the prior art requires high professional skills. Both manual teaching and offline programming increase labor costs and training difficulties, and it is difficult to quickly adjust in the face of changes in the actual situation of the casting.

[0004] To this end, the present invention provides an artificial intelligence-based autonomous decision-making auxiliary system for a casting cleaning robot. Summary of the invention

[0005] The purpose of the present invention is to provide an artificial intelligence-based casting cleaning robot autonomous decision-making assistance system to solve the above-mentioned background problems.

[0006] The purpose of the present invention can be achieved through the following technical solutions: An autonomous decision-making auxiliary system for a casting cleaning robot based on artificial intelligence, comprising: Parameter acquisition module: obtain the size and concavity information of the sub-block, calculate and obtain the regional flatness Pz, the regional flatness Pz ≤ the regional flatness threshold Pzy, and mark the sub-block as a priority processing block; Parameter analysis module: obtain the difference between the regional flatness Pz and the threshold Pzy, calculate the flatness difference proximity ratio Pj, calculate the concave and convex density of the sub-blocks in the priority processing block, and obtain the block concave-convex ratio. According to the mean of the sub-block concave-convex ratio, calculate the block concave-convex difference ratio Czb. Based on the flatness difference proximity ratio Pj and the block concave-convex difference ratio Czb, calculate the priority cleaning coefficient Yx. If the priority cleaning coefficient Yx ≥ the priority cleaning coefficient threshold Yxz, mark the priority processing block as a block to be cleaned. Path planning module: Based on the blocks to be cleaned, the clustering algorithm and A* algorithm in the field of artificial intelligence are used to build an AI cleaning path model to calculate the cleaning path of the blocks to be cleaned; Decision execution module: obtain the cleaning time of the casting cleaning robot, further calculate the cleaning time ratio Sj, obtain the path cleaning ratio Ql based on the regional flatness Pz of the cleaning path, calculate the decision path coefficient Jc based on the cleaning time ratio Sj and the path cleaning ratio Ql, and select the optimal cleaning path to control the casting cleaning robot to clean the casting surface according to the optimal cleaning processing path.

[0007] As a further technical solution of the present invention: the method for obtaining the regional flatness Pz is: The regional flatness Pz is calculated based on the convex number ratio Tq, the concave number ratio Ax, and the concave-convex area ratio Mj; By formula: , obtain the regional flatness Pz, where a1, a2, and a3 are preset proportional coefficients.

[0008] As a further technical solution of the present invention: the method for obtaining the protrusion quantity ratio Tq and the depression quantity ratio Ax is as follows: Acquire a detection model of the casting, divide the surface of the detection model of the casting into multiple blocks, and obtain sub-blocks; Obtain the number of convexities and concavities in the sub-block, and sum the number of convexities and concavities in the sub-block to obtain the sum of the number of concavities; The number of convexities and concavities are respectively processed by ratio processing with the sum of the number of concavities and convexities to obtain the ratio of the number of convexities and the ratio of the number of concavities in the sub-block; The ratio of the number of convexities in the sub-block is marked as Tq, and the ratio of the number of concavities is marked as Ax; The concave-convex area ratio Mj is obtained as follows: Obtain the areas of the convexities and concavities in the sub-block, and sum the areas of the convexities and concavities in the sub-block to obtain the sum of the concave and convex areas; The sum of the concave and convex areas is ratioed to the area of ​​the sub-block to obtain the concave and convex area ratio, which is marked as Mj.

[0009] As a further technical solution of the present invention: the method for obtaining the priority processing block is: Compare the regional flatness Pz with the regional flatness threshold Pzy; If the regional flatness Pz ≤ the regional flatness threshold Pzy, the sub-block is marked as a priority processing block.

[0010] As a further technical solution of the present invention: the priority cleaning coefficient Yx is obtained as follows: Based on the leveling difference approach ratio Pj and the block concave-convex difference ratio Czb, the priority cleaning coefficient Yx is calculated; By formula: , calculate the priority cleaning coefficient Yx, where b1 and b2 are preset proportional coefficients; The method for obtaining the block to be cleaned is as follows: Compare the priority cleaning coefficient Yx of the priority cleaning block with the priority cleaning coefficient threshold Yxz; If the priority cleaning coefficient Yx≥priority cleaning coefficient threshold Yxz, the priority processing block is marked as a block to be cleaned.

[0011] As a further technical solution of the present invention: the method for obtaining the leveling difference proximity ratio Pj is: Obtain the difference between the regional flatness Pz of the priority processing block and the regional flatness threshold Pzy to obtain the flatness difference; The flatness difference is processed by ratio processing with the regional flatness threshold Pzy to obtain the flatness difference proximity ratio, and the flatness difference proximity ratio is marked as Pj.

[0012] As a further technical solution of the present invention: the block concave-convex difference ratio Czb is obtained as follows: Based on the priority processing block, the number of depressions in the sub-block is processed by ratio with the area of ​​the sub-block to obtain the block depression density; The number of protrusions in the sub-block is processed by ratio with the area of ​​the sub-block to obtain the block protrusion density; The block concave density and the block convex density are processed by ratio to obtain the block concave-convex ratio; Obtain the block concavo-convex ratios of all sub-blocks, sum and average the block concavo-convex ratios of all sub-blocks, and obtain the mean value of the concavo-convex ratios; Subtract the block concavo-convex ratio from the mean value of the concavo-convex ratio and take the absolute value to obtain the block concavo-convex ratio difference; The block concave-convex ratio difference is processed with the concave-convex ratio mean to obtain the block concave-convex difference ratio, and the block concave-convex difference ratio is marked as Czb.

[0013] As a further technical solution of the present invention: the AI ​​cleaning path model is obtained in the following manner: S1: Perform cluster analysis on the blocks to be cleaned, use the block concave-convex difference ratio Czb and the flat difference ratio Pj as parameter values, and use the K-Means algorithm to divide the blocks with similar parameter values ​​into one cluster; S2: Calculate the distance between each block to be cleaned and each cluster center, assign each block to be cleaned to the cluster to which the nearest cluster center belongs, and update the distance center; S3: Repeat S2 until the position change of the cluster center reaches the set number of iterations 100, and take the block with the smallest leveling difference ratio Pj as the starting block; S4: For each cluster, use the A* algorithm in path planning, starting from the cluster starting block, until all the blocks to be cleaned in the cluster have been visited, and find a cleaning path in each cluster.

[0014] As a further technical solution of the present invention: the decision path coefficient Jc is obtained as follows: Based on the cleaning time ratio Sj and the path cleaning ratio Ql, the decision path coefficient Jc is calculated; By formula: , calculate the decision path coefficient Jc, where b1 and b2 are preset proportional coefficients; The optimal cleaning path is obtained as follows: Obtain the decision path coefficients Jc of all cleaning paths of the block to be cleaned, sort the decision path coefficients Jc in descending order, select the maximum value of the decision path coefficient Jc, and obtain the optimal cleaning path; Use a casting cleaning robot to clean the casting surface according to the optimal cleaning path.

[0015] As a further technical solution of the present invention: the cleaning time ratio Sj and the path cleaning ratio Ql are obtained as follows: Based on the AI ​​cleaning path model, the cleaning path within each cluster is obtained to obtain the path to be cleaned; Calculate the length of the path to be cleaned and obtain the cleaning path length; Acquire the cleaning speed of the casting cleaning robot to obtain the casting cleaning speed; The cleaning path length is ratioed to the casting cleaning speed to obtain the casting cleaning time; The casting cleaning time is processed by ratio processing with the casting cleaning time threshold to obtain the cleaning time ratio, and the cleaning time ratio is marked as Sj; Obtain the regional flatness Pz of the cleaning path of the block to be cleaned, calculate the sum of the regional flatness Pz on the cleaning path, and obtain the sum of the path flatness; Obtain the regional flatness Pz of the block to be cleaned, calculate the sum of the regional flatness Pz on the block to be cleaned, and obtain the sum of the block flatness; The sum of the path flatness and the sum of the block flatness are ratio processed to obtain the path cleaning ratio, which is marked as Ql.

[0016] Beneficial effects of the present invention: The parameter acquisition module is used to obtain sub-block information to calculate the flatness of the area and distinguish between priority processing blocks and blocks to be processed. The parameter calculation module further calculates the priority cleaning coefficient based on the flatness difference proximity ratio and the block convexity difference ratio to determine the blocks to be cleaned. The path planning module plans the cleaning path by building an AI cleaning path model. Compared with the traditional method that relies on manual experience or simple programming, it greatly improves the rationality and accuracy of path planning, effectively shortens the length of the cleaning path, and improves the cleaning efficiency. By calculating the cleaning time ratio and the path cleaning ratio, the decision path coefficient Jc is obtained, thereby determining the optimal cleaning path, improving the cleaning efficiency, and making the cleaning quality of the casting surface more uniform. At the same time, the same intelligent decision-making process can be used for the blocks to be processed, further enhancing the system's adaptability to different block conditions, overcoming the problem of insufficient flexibility of existing technologies when facing casting changes, and effectively improving the overall quality and intelligence level of casting cleaning. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The present invention will be further described below in conjunction with the accompanying drawings.

[0018] Figure 1 This is a module diagram of an autonomous decision-making auxiliary system of a casting cleaning robot based on artificial intelligence of the present invention; Figure 2 It is a flow chart of building the AI ​​cleaning path model in the present invention. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] Example 1, please refer to Figure 1 As shown, the present invention is an autonomous decision-making auxiliary system for a casting cleaning robot based on artificial intelligence, comprising: Parameter acquisition module: obtain the concavity and convexity information of the sub-block, calculate and obtain the regional flatness Pz, the regional flatness Pz ≤ the regional flatness threshold Pzy, and mark the sub-block as a priority processing block; Build a casting inspection model based on visual sensors and 3D scanning technology; Compare the casting inspection model with the casting reference model to obtain the surface concavity and convexity data of the casting inspection model; Divide the surface of the casting inspection model into multiple blocks to obtain sub-blocks; Obtain the number of convexities and concavities in the sub-block, and sum the number of convexities and concavities in the sub-block to obtain the sum of the number of concavities; The number of convexities and concavities are respectively processed by ratio processing with the sum of the number of concavities and convexities to obtain the ratio of the number of convexities and the ratio of the number of concavities in the sub-block; The ratio of the number of convexities in the sub-block is marked as Tq, and the ratio of the number of concavities is marked as Ax; Obtain the areas of the convexities and concavities in the sub-block, and sum the areas of the convexities and concavities in the sub-block to obtain the sum of the concave and convex areas; The sum of the concave and convex areas is processed by ratio with the area of ​​the sub-block to obtain the concave and convex area ratio, which is marked as Mj; The regional flatness Pz is calculated based on the convex number ratio Tq, the concave number ratio Ax, and the concave-convex area ratio Mj; By formula: , obtain the regional flatness Pz, where a1, a2, and a3 are preset proportional coefficients, , represents the logarithmic function with base e; Compare the regional flatness Pz with the regional flatness threshold Pzy; If the regional flatness Pz ≤ the regional flatness threshold Pzy, it indicates that the flatness of the casting sub-block is poor, and the sub-block is marked as a priority processing block; If the regional flatness Pz>regional flatness threshold Pzy, it indicates that the flatness of the casting sub-block is good, and the sub-block is marked as a block to be processed; Parameter analysis module: obtain the difference between the regional flatness Pz of the priority processing block and the threshold Pzy, calculate the flatness difference proximity ratio Pj, calculate the concave and convex density of the sub-blocks in the priority processing block, and obtain the block concave-convex ratio. According to the average of the concave-convex ratios of all sub-blocks, calculate the block concave-convex difference ratio Czb. Based on the flatness difference proximity ratio Pj and the block concave-convex difference ratio Czb, calculate the priority cleaning coefficient Yx. If the priority cleaning coefficient Yx ≥ the priority cleaning coefficient threshold Yxz, mark the priority processing block as a block to be cleaned; Obtain the difference between the regional flatness Pz of the priority processing block and the regional flatness threshold Pzy to obtain the flatness difference; The flatness difference is processed by ratio with the regional flatness threshold Pzy to obtain the flatness difference approximation ratio, and the flatness difference approximation ratio is marked as Pj; Based on the priority processing blocks, the number of depressions and the number of protrusions in the sub-blocks are ratio-processed with the sub-block area to obtain the block depression density and block protrusion density respectively; The block concave density and the block convex density are processed by ratio to obtain the block concave-convex ratio; Obtain the block concavo-convex ratios of all sub-blocks, sum and average the block concavo-convex ratios of all sub-blocks, and obtain the mean value of the concavo-convex ratios; Subtract the block concavo-convex ratio from the mean value of the concavo-convex ratio and take the absolute value to obtain the block concavo-convex ratio difference; The block concave-convex ratio difference is processed by ratio with the concave-convex ratio mean to obtain the block concave-convex difference ratio, and the block concave-convex difference ratio is marked as Czb; It should be noted that the block concave-convex difference ratio Czb reflects the concave-convex characteristics of a specific sub-block on the casting surface relative to the overall average concave-convex degree. If the block concave-convex difference ratio Czb is larger, it indicates that there is a large difference between the concave-convex ratio of the sub-block and the overall average concave-convex ratio, and the sub-block has potential quality risks. Based on the leveling difference approach ratio Pj and the block concave-convex difference ratio Czb, the priority cleaning coefficient Yx is calculated; By formula: , calculate the priority cleaning coefficient Yx, where b1 and b2 are preset proportional coefficients; Compare the priority cleaning coefficient Yx of the priority cleaning block with the priority cleaning coefficient threshold Yxz; If the priority cleaning coefficient Yx ≥ the priority cleaning coefficient threshold Yxz, the priority processing block is marked as a block to be cleaned; If the priority cleaning coefficient Yx is less than the priority cleaning coefficient threshold Yxz, no processing is performed; The technical solution of this embodiment is: obtain the concavity and convexity information of the sub-block, calculate and obtain the regional flatness Pz, the regional flatness Pz≤regional flatness threshold Pzy, mark the sub-block as a priority processing block, obtain the difference between the regional flatness Pz of the priority processing block and the threshold Pzy, calculate the flatness difference proximity ratio Pj, calculate the block concavity and convexity ratio of the sub-block in the priority processing block, calculate it with the average of the concavity and convexity ratios of all sub-blocks, and obtain the block concavity and convexity difference ratio Czb, and then calculate the priority cleaning coefficient Yx based on the flatness difference proximity ratio Pj and the block concavity and convexity difference ratio Czb, if the priority cleaning coefficient Yx≥priority cleaning coefficient threshold Yxz, mark the priority processing block as a block to be cleaned.

[0021] Embodiment 2, based on embodiment 1, an artificial intelligence-based casting cleaning robot autonomous decision-making assistance system further includes: Path planning module: Based on the blocks to be cleaned, the clustering algorithm and A* algorithm in the field of artificial intelligence are used to build an AI cleaning path model to calculate the cleaning path of the blocks to be cleaned; Specifically, a clustering algorithm in the field of artificial intelligence is used to build an AI cleaning path model; like Figure 2 As shown, the AI ​​cleaning path model is constructed as follows: S1: cluster analysis is performed on the blocks to be cleaned. Based on the block concave-convex difference ratio Czb and the flat difference ratio Pj of the blocks to be cleaned, the K-Means algorithm is used to divide the blocks with similar parameter values ​​into a cluster, and the number of clusters K is determined to be 5; S2: Randomly initialize 5 cluster centers, calculate the distance between each block to be cleaned and each cluster center, assign each block to be cleaned to the cluster to which the nearest cluster center belongs, and update the distance center; S3: Repeat S2 until the position change of the cluster center reaches the set number of iterations 100, and a stable cluster is obtained. In each cluster, all the blocks to be cleaned are traversed, and the block with the smallest leveling difference ratio Pj is used as the starting block; S4: For each cluster, use the A* algorithm in path planning, starting from the cluster starting block, until all the blocks to be cleaned in the cluster have been visited, and find a cleaning path in each cluster; It should be noted that the cleaning path obtained through the AI ​​cleaning path model is not unique; Decision execution module: obtain the cleaning time of the casting cleaning robot, further calculate the cleaning time ratio Sj, obtain the path cleaning ratio Ql based on the regional flatness Pz of the cleaning path, calculate the decision path coefficient Jc based on the cleaning time ratio Sj and the path cleaning ratio Ql, and select the optimal cleaning path to control the casting cleaning robot to clean the casting surface according to the optimal cleaning treatment path; Based on the AI ​​cleaning path model, the cleaning path within each cluster is obtained to obtain the path to be cleaned; Calculate the length of the path to be cleaned and obtain the cleaning path length; Acquire the cleaning speed of the casting cleaning robot to obtain the casting cleaning speed; The cleaning path length is ratioed to the casting cleaning speed to obtain the casting cleaning time; The casting cleaning time is processed by ratio processing with the casting cleaning time threshold to obtain the cleaning time ratio, and the cleaning time ratio is marked as Sj; Obtain the regional flatness Pz of the cleaning path of the block to be cleaned, calculate the sum of the regional flatness Pz on the cleaning path, and obtain the sum of the path flatness; Obtain the regional flatness Pz of the block to be cleaned, calculate the sum of the regional flatness Pz on the block to be cleaned, and obtain the sum of the block flatness; The path cleaning ratio is obtained by performing a ratio process on the sum of the path flatness and the sum of the block flatness. The path cleaning ratio is labeled as Ql; Based on the cleaning time ratio Sj and the path cleaning ratio Ql, the decision path coefficient Jc is calculated; By formula: , calculate the decision path coefficient Jc, where b1 and b2 are preset proportional coefficients; Obtain the decision path coefficients Jc of all cleaning paths of the block to be cleaned, sort the decision path coefficients Jc in descending order, select the maximum value of the decision path coefficient Jc, and obtain the optimal cleaning path; Use a casting cleaning robot to clean the casting surface according to the optimal cleaning path; For the blocks to be processed, the AI ​​cleaning path model is used to calculate the cleaning path and the decision path coefficient Jc of the blocks to be processed is calculated; The optimal cleaning path of the block to be processed is calculated, and the surface of the casting is cleaned according to the optimal cleaning path; The technical solution of this embodiment is: based on the block to be cleaned, the clustering algorithm and A* algorithm in the field of artificial intelligence are used to construct an AI cleaning path model, the cleaning path of the block to be cleaned is calculated, the cleaning time of the casting cleaning robot is obtained, and the cleaning time ratio Sj is further calculated. Based on the regional flatness Pz of the cleaning path, the path cleaning ratio Ql is obtained, based on the cleaning time ratio Sj and the path cleaning ratio Ql, the decision path coefficient Jc is calculated, and the optimal cleaning path is selected to control the casting cleaning robot to clean the casting surface according to the optimal cleaning processing path.

[0022] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. An autonomous decision-making assistance system for casting cleaning robots based on artificial intelligence, characterized in that: include: Parameter acquisition module: obtain the size and concavity information of the sub-block, calculate and obtain the regional flatness Pz, if the regional flatness Pz ≤ the regional flatness threshold Pzy, mark the sub-block as a priority processing block; Parameter analysis module: obtain the difference between the regional flatness Pz and the threshold Pzy, calculate the flatness difference approximation ratio Pj, calculate the concave and convex density of the sub-blocks in the priority processing block, and obtain the block concave-convex ratio. According to the mean of the sub-block concave-convex ratio, calculate the block concave-convex difference ratio Czb. Based on the flatness difference approximation ratio Pj and the block concave-convex difference ratio Czb, calculate the priority cleaning coefficient Yx. If the priority cleaning coefficient Yx ≥ the priority cleaning coefficient threshold Yxz, mark the priority processing block as a block to be cleaned; Path planning module: Based on the blocks to be cleaned, the clustering algorithm and A* algorithm in the field of artificial intelligence are used to build an AI cleaning path model to calculate the cleaning path of the blocks to be cleaned; Decision execution module: obtain the cleaning time of the casting cleaning robot, calculate the cleaning time ratio Sj, obtain the path cleaning ratio Ql based on the regional flatness Pz of the cleaning path, calculate the decision path coefficient Jc based on the cleaning time ratio Sj and the path cleaning ratio Ql, and select the optimal cleaning path to control the casting cleaning robot to clean the casting surface according to the optimal cleaning processing path.

2. The autonomous decision-making support system for casting cleaning robots based on artificial intelligence according to claim 1, characterized in that: The method for obtaining the regional flatness Pz is: The regional flatness Pz is calculated based on the convex number ratio Tq, the concave number ratio Ax, and the concave-convex area ratio Mj; By formula: , obtain the regional flatness Pz, where a1, a2, and a3 are preset proportional coefficients.

3. The autonomous decision-making support system for casting cleaning robots based on artificial intelligence according to claim 2, characterized in that: The method for obtaining the protrusion quantity ratio Tq and the depression quantity ratio Ax is as follows: Acquire a detection model of the casting, divide the surface of the detection model of the casting into multiple blocks, and obtain sub-blocks; Obtain the number of convexities and concavities in the sub-block, and sum the number of convexities and concavities in the sub-block to obtain the sum of the number of concavities; The number of convexities and concavities are respectively processed by ratio processing with the sum of the number of concavities and convexities to obtain the ratio of the number of convexities and the ratio of the number of concavities in the sub-block; The ratio of the number of convexities in the sub-block is marked as Tq, and the ratio of the number of concavities is marked as Ax; The concave-convex area ratio Mj is obtained as follows: Obtain the areas of the convexities and concavities in the sub-block, and sum the areas of the convexities and concavities in the sub-block to obtain the sum of the concave and convex areas; The sum of the concave and convex areas is ratioed to the area of ​​the sub-block to obtain the concave and convex area ratio, which is marked as Mj.

4. The autonomous decision-making support system for casting cleaning robots based on artificial intelligence according to claim 2, characterized in that: The method for obtaining the priority processing block is as follows: Compare the regional flatness Pz with the regional flatness threshold Pzy; If the regional flatness Pz ≤ the regional flatness threshold Pzy, the sub-block is marked as a priority processing block.

5. The autonomous decision-making support system for casting cleaning robots based on artificial intelligence according to claim 1, characterized in that: The priority cleaning coefficient Yx is obtained as follows: Based on the leveling difference approach ratio Pj and the block concave-convex difference ratio Czb, the priority cleaning coefficient Yx is calculated; By formula: , calculate the priority cleaning coefficient Yx, where b1 and b2 are preset proportional coefficients; The method for obtaining the block to be cleaned is: Compare the priority cleaning coefficient Yx of the priority cleaning block with the priority cleaning coefficient threshold Yxz; If the priority cleaning coefficient Yx≥priority cleaning coefficient threshold Yxz, the priority processing block is marked as a block to be cleaned.

6. The autonomous decision-making support system for casting cleaning robots based on artificial intelligence according to claim 5 is characterized in that: The method for obtaining the leveling difference approach ratio Pj is as follows: Obtain the difference between the regional flatness Pz of the priority processing block and the regional flatness threshold Pzy to obtain the flatness difference; The flatness difference is processed by ratio processing with the regional flatness threshold Pzy to obtain the flatness difference proximity ratio, and the flatness difference proximity ratio is marked as Pj.

7. The autonomous decision-making support system for casting cleaning robots based on artificial intelligence according to claim 5 is characterized in that: The block concave-convex difference ratio Czb is obtained as follows: Based on the priority processing blocks, the number of depressions and the number of protrusions in the sub-blocks are ratio-processed with the sub-block area to obtain the block depression density and block protrusion density respectively; The block concave density and the block convex density are processed by ratio to obtain the block concave-convex ratio; Obtain the block concavo-convex ratios of all sub-blocks, sum and average the block concavo-convex ratios of all sub-blocks, and obtain the mean value of the concavo-convex ratios; Subtract the block concavo-convex ratio from the mean value of the concavo-convex ratio and take the absolute value to obtain the block concavo-convex ratio difference; The block concave-convex ratio difference is processed with the concave-convex ratio mean to obtain the block concave-convex difference ratio, and the block concave-convex difference ratio is marked as Czb.

8. The autonomous decision-making support system for casting cleaning robots based on artificial intelligence according to claim 1, characterized in that: The AI ​​cleaning path model is obtained as follows: S1: Perform cluster analysis on the blocks to be cleaned, use the block concave-convex difference ratio Czb and the flat difference ratio Pj as parameter values, and use the K-Means algorithm to divide the blocks with similar parameter values ​​into one cluster; S2: Calculate the distance between each block to be cleaned and each cluster center, assign each block to be cleaned to the cluster to which the nearest cluster center belongs, and update the distance center; S3: Repeat S2 until the position change of the cluster center reaches the set number of iterations 100, and take the block with the smallest leveling difference ratio Pj as the starting block; S4: For each cluster, use the A* algorithm in path planning, starting from the cluster starting block, until all the blocks to be cleaned in the cluster have been visited, and find a cleaning path in each cluster.

9. The autonomous decision-making support system for casting cleaning robots based on artificial intelligence according to claim 1, characterized in that: The decision path coefficient Jc is obtained as follows: Based on the cleaning time ratio Sj and the path cleaning ratio Ql, the decision path coefficient Jc is calculated; By formula: , calculate the decision path coefficient Jc, where b1 and b2 are preset proportional coefficients; The optimal cleaning path is obtained as follows: Obtain the decision path coefficients Jc of all cleaning paths of the block to be cleaned, sort the decision path coefficients Jc in descending order, select the maximum value of the decision path coefficient Jc, and obtain the optimal cleaning path; Use a casting cleaning robot to clean the casting surface according to the optimal cleaning path.

10. The autonomous decision-making support system for casting cleaning robots based on artificial intelligence according to claim 1, characterized in that: The cleaning time ratio Sj and the path cleaning ratio Ql are obtained as follows: Based on the AI ​​cleaning path model, the cleaning path within each cluster is obtained to obtain the path to be cleaned; Calculate the length of the path to be cleaned and obtain the cleaning path length; Acquire the cleaning speed of the casting cleaning robot to obtain the casting cleaning speed; The cleaning path length is ratioed to the casting cleaning speed to obtain the casting cleaning time; The casting cleaning time is processed by ratio processing with the casting cleaning time threshold to obtain the cleaning time ratio, and the cleaning time ratio is marked as Sj; Obtain the regional flatness Pz of the cleaning path of the block to be cleaned, calculate the sum of the regional flatness Pz on the cleaning path, and obtain the sum of the path flatness; Obtain the regional flatness Pz of the block to be cleaned, calculate the sum of the regional flatness Pz on the block to be cleaned, and obtain the sum of the block flatness; The sum of the path flatness and the sum of the block flatness are ratio processed to obtain the path cleaning ratio, which is marked as Ql.

Images