Grinding and scraping robot and control method based on learning algorithm

Abstract

The present application relates to the field of robot technology, and provides a grinding and scraping robot and control method based on a learning algorithm, the method comprising: constructing a three-dimensional wall model of a wall to be processed, and generating an initial grinding and scraping scheme of the three-dimensional wall model based on a preset initial grinding and scraping scheme generation method; judging whether the initial grinding and scraping scheme can effectively grind and scrape the wall to be processed; if not, reinforcing the initial grinding and scraping scheme based on a preset reinforcement learning algorithm to obtain a target grinding and scraping scheme; and controlling the grinding and scraping robot to grind and scrape the wall to be processed based on the target grinding and scraping scheme. This method can achieve refined control of the wall and achieve refined grinding of complex walls.

Description

Technical Field

[0001] The present application relates to the field of robotics technology, and in particular to a grinding and scraping robot based on a learning algorithm and a control method thereof. Background Art

[0002] In the process of building decoration and construction, wall polishing is a key step to ensure the smoothness and beauty of the wall surface. Traditional wall polishing mostly relies on manual operation, which is not only inefficient but also labor-intensive, which can easily lead to operator fatigue and cause safety accidents. Although the existing automated polishing equipment can improve polishing efficiency and safety, it has a low level of intelligence and is difficult to achieve fine polishing control of the wall. Summary of the invention

[0003] The present application provides a grinding and scraping robot and a control method based on a learning algorithm to solve the problems raised by the above background technology.

[0004] In a first aspect, the present application provides a control method for a grinding and scraping robot based on a learning algorithm, comprising:

[0005] Constructing a three-dimensional wall model of the wall to be processed, and generating an initial grinding and scraping scheme of the three-dimensional wall model based on a preset initial grinding and scraping scheme generation method;

[0006] Determine whether the initial grinding and scraping scheme can effectively grind and scrape the wall surface to be processed;

[0007] If not, the initial grinding and scraping scheme is subjected to reinforcement learning based on a preset reinforcement learning algorithm to obtain a target grinding and scraping scheme;

[0008] Based on the target grinding and scraping scheme, the grinding and scraping robot is controlled to grind and scrape the wall surface to be processed.

[0009] In a possible implementation, the step of constructing a three-dimensional wall model of the wall to be processed and generating an initial grinding and scraping scheme of the three-dimensional wall model based on a preset initial grinding and scraping scheme generating method includes:

[0010] Acquire image information and first point cloud data information of a wall surface to be processed, and generate a three-dimensional wall model of the wall surface to be processed based on the image information and the first point cloud data information;

[0011] Control the virtual grinding and scraping robot to perform grinding and scraping processing on the three-dimensional wall model, and generate the initial grinding and scraping plan during the process of the virtual grinding and scraping robot performing grinding and scraping processing on the three-dimensional wall model; wherein the virtual grinding and scraping robot matches the grinding and scraping robot.

[0012] In a possible implementation, the acquiring image information and first point cloud data information of a wall to be processed, and generating a three-dimensional wall model of the wall to be processed based on the image information and the first point cloud data information, includes:

[0013] The image information is acquired by a camera device preset on the polishing and scraping robot, and the first point cloud data information is acquired by a laser scanner preset on the polishing and scraping robot;

[0014] The geometric structure of the wall to be processed is constructed based on the first point cloud data information, and the geometric structure is rendered based on the image information to obtain the three-dimensional model of the wall.

[0015] In a possible implementation, the controlling the virtual grinding and scraping robot to perform grinding and scraping processing on the three-dimensional wall model, and generating the initial grinding and scraping scheme during the process of the virtual grinding and scraping robot performing grinding and scraping processing on the three-dimensional wall model, includes:

[0016] Determining a polishing area of ​​the three-dimensional wall model;

[0017] Constructing a reference surface of the three-dimensional wall model directly in front of the three-dimensional wall model based on the outer contour of the three-dimensional wall model;

[0018] Obtaining first distance values ​​between each first geometric point in the polishing area and the reference surface;

[0019] Counting the number of each of the first distance values ​​respectively, and determining the surface where the first geometric points corresponding to the first distance values ​​with the largest number are located as the reference surface of the three-dimensional wall model;

[0020] The reference plane is marked in the three-dimensional model of the wall, and the virtual grinding and scraping robot is controlled to perform grinding and scraping on the grinding area based on the reference plane. In the process of the virtual grinding and scraping robot performing grinding and scraping on the grinding area, the operating parameter information of the grinding and scraping robot is collected in real time to generate the initial grinding and scraping plan.

[0021] In a possible implementation, the determining whether the initial grinding and scraping scheme can effectively grind and scrape the wall surface to be processed includes:

[0022] Acquire second point cloud data information of the polished area after polishing and scraping, and acquire second distance values ​​between each second geometric point of the second point cloud data information and the reference surface;

[0023] Obtaining an initial standard deviation between each of the second distance values, and comparing the initial standard deviation with a preset standard deviation;

[0024] If the initial standard deviation is less than the preset standard deviation, it is determined that the initial grinding and scraping scheme can effectively grind and scrape the wall surface to be processed;

[0025] If the initial standard deviation is not less than the preset standard deviation, it is determined that the initial grinding and scraping scheme cannot effectively grind and scrape the wall surface to be processed.

[0026] In a possible implementation, the initial grinding and scraping scheme is subjected to reinforcement learning based on a preset reinforcement learning algorithm to obtain a target grinding and scraping scheme, including:

[0027] Inputting the three-dimensional wall model after the grinding and scraping process into a preset wall defect detection model to obtain defect information of the three-dimensional wall model after the grinding and scraping process; wherein the wall defect detection model is a pre-trained neural network model;

[0028] Generating a repair plan for the three-dimensional model of the wall after grinding and scraping based on the defect information;

[0029] Optimizing the initial grinding and scraping scheme based on the repair scheme to obtain an intermediate grinding and scraping scheme;

[0030] Controlling the virtual grinding and scraping robot to perform grinding and scraping processing on the grinding area based on the intermediate grinding and scraping scheme, and obtaining third point cloud data information of the grinding area after the grinding and scraping processing, and obtaining third distance values ​​between each third geometric point of the third point cloud data information and the reference plane;

[0031] Obtaining an intermediate standard deviation between each of the third distance values, and comparing the intermediate standard deviation with a preset standard deviation;

[0032] If the intermediate standard deviation is not less than the preset standard deviation, the three-dimensional wall model after grinding and scraping is input into the preset wall defect detection model in an iterative loop to obtain the defect information of the three-dimensional wall model after grinding and scraping and subsequent steps until the intermediate standard deviation is less than the preset standard deviation to obtain the target grinding and scraping plan.

[0033] In a possible implementation, the training method of the wall defect detection model includes:

[0034] Obtaining a training wall three-dimensional model and label information corresponding to the training wall three-dimensional model;

[0035] The training wall three-dimensional model and the label information are used as a training pair and input into a neural network model for training;

[0036] Obtaining a test wall three-dimensional model and standard label information corresponding to the test wall three-dimensional model;

[0037] Inputting the test wall three-dimensional model into the trained neural network model to obtain predicted label information of the test wall three-dimensional model;

[0038] Obtaining a loss value between the standard label information and the predicted label information;

[0039] comparing the loss value with a preset loss value;

[0040] If the loss value is greater than the preset loss value, reversely adjusting the model parameters of the trained neural network model based on the loss value;

[0041] The above training method is iterated until the loss value is no greater than the preset loss value, and the trained neural network model is used as the wall defect detection model.

[0042] In a second aspect, the present application provides a grinding and scraping robot based on a learning algorithm, comprising:

[0043] A first generation module is used to construct a three-dimensional wall model of the wall to be processed, and generate an initial grinding and scraping scheme of the three-dimensional wall model based on a preset initial grinding and scraping scheme generation method;

[0044] A judgment module, used to judge whether the initial grinding and scraping scheme can effectively grind and scrape the wall surface to be processed;

[0045] The second generation module is used for performing reinforcement learning on the initial grinding and scraping scheme based on a preset reinforcement learning algorithm to obtain a target grinding and scraping scheme if the initial grinding and scraping scheme cannot effectively grind and scrape the wall surface to be treated;

[0046] The control module is used to control the grinding and scraping robot to grind and scrape the wall surface to be processed based on the target grinding and scraping scheme.

[0047] The present application provides a grinding and scraping robot and control method based on a learning algorithm, the method comprising: constructing a three-dimensional wall model of a wall to be processed, and generating an initial grinding and scraping scheme of the three-dimensional wall model based on a preset initial grinding and scraping scheme generation method; judging whether the initial grinding and scraping scheme can effectively grind and scrape the wall to be processed; if not, reinforcing the initial grinding and scraping scheme based on a preset reinforcement learning algorithm to obtain a target grinding and scraping scheme; and controlling the grinding and scraping robot to grind and scrape the wall to be processed based on the target grinding and scraping scheme. This method, on the one hand, can realize the grinding and scraping of complex walls and improve the intelligent level of wall grinding and scraping; on the other hand, it can realize the refined grinding and scraping of walls and improve the aesthetics of the walls. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying any creative work.

[0049] Figure 1 A schematic flow chart of a control method for a grinding and scraping robot based on a learning algorithm provided in an embodiment of the present application;

[0050] Figure 2 A schematic block diagram of the structure of a grinding and scraping robot based on a learning algorithm provided in an embodiment of the present application;

[0051] Figure 3 A schematic block diagram of the structure of a terminal device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0052] 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 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.

[0053] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.

[0054] It should also be understood that the terms used in this application specification are only for the purpose of describing specific embodiments and are not intended to limit the application. As used in this application specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0055] It should be further understood that the term “and / or” used in the specification and appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0056] In conjunction with the accompanying drawings, some embodiments of the present application are described in detail below. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0057] See also Figure 1 , Figure 1 A flow chart of a control method for a grinding and scraping robot based on a learning algorithm provided in an embodiment of the present application is shown in FIG. Figure 1 As shown, the grinding and scraping robot control method based on learning algorithm provided in the embodiment of the present application includes steps S1 to S4.

[0058] Step S1: construct a three-dimensional wall model of the wall to be processed, and generate an initial grinding and scraping plan for the three-dimensional wall model based on a preset initial grinding and scraping plan generation method.

[0059] It should be noted that the executor of this embodiment is a grinding and scraping robot based on a learning algorithm.

[0060] Specifically, step S1 includes the following steps:

[0061] Acquire image information and first point cloud data information of a wall surface to be processed, and generate a three-dimensional wall model of the wall surface to be processed based on the image information and the first point cloud data information;

[0062] Control the virtual grinding and scraping robot to perform grinding and scraping processing on the three-dimensional wall model, and generate the initial grinding and scraping plan during the process of the virtual grinding and scraping robot performing grinding and scraping processing on the three-dimensional wall model; wherein the virtual grinding and scraping robot matches the grinding and scraping robot.

[0063] The step of acquiring the image information and the first point cloud data information of the wall to be processed, and generating a three-dimensional wall model of the wall to be processed based on the image information and the first point cloud data information, comprises the following steps:

[0064] The image information is acquired by a camera device preset on the polishing and scraping robot, and the first point cloud data information is acquired by a laser scanner preset on the polishing and scraping robot;

[0065] The geometric structure of the wall to be processed is constructed based on the first point cloud data information, and the geometric structure is rendered based on the image information to obtain the three-dimensional model of the wall.

[0066] It should be noted that when the laser scanner scans the wall to be treated, it generates first three-dimensional coordinate points of each first geometric point of the wall to be treated, and each of the first three-dimensional coordinate points constitutes the first point cloud data information. It can be understood that the coordinate systems corresponding to each of the first three-dimensional coordinate points are consistent. After the grinding and scraping robot obtains the first point cloud data information, it constructs a coordinate system corresponding to the first point cloud data information, and marks the first geometric points corresponding to each of the first three-dimensional coordinate points on the coordinate system to obtain the geometric structure of the wall to be treated.

[0067] It can be understood that the method provided in this embodiment, by rendering the geometric structure based on the image information, obtains the three-dimensional model of the wall, and can prevent the virtual robot from colliding with obstacles while controlling the virtual grinding and scraping robot to perform grinding and scraping on the grinding area. The obstacles include but are not limited to switches, sockets and wall decorations.

[0068] The step of controlling the virtual grinding and scraping robot to perform grinding and scraping processing on the three-dimensional wall model, and generating the initial grinding and scraping scheme in the process of the virtual grinding and scraping robot performing grinding and scraping processing on the three-dimensional wall model, comprises the following steps:

[0069] Determine the polishing area of ​​the three-dimensional wall model; specifically, determine the area outside the area where the virtual obstacle on the three-dimensional wall model is located as the polishing area; it can be understood that the virtual obstacle corresponds to the obstacle on the wall to be processed;

[0070] Constructing a reference surface of the three-dimensional wall model in front of the three-dimensional wall model based on the external contour of the three-dimensional wall model; wherein the external contour refers to the contour of the wall;

[0071] Obtaining first distance values ​​between each first geometric point in the polishing area and the reference surface; wherein, for each first geometric point, the first geometric point corresponds to a unique first target point on the reference surface, the tangent plane of the first geometric point is parallel to the tangent plane of the first target point, the line connecting the first geometric point and the first target point is perpendicular to the tangent plane of the first geometric point, and the distance value between the first geometric point and the first target point is the first distance value corresponding to the first geometric point;

[0072] Counting the number of each of the first distance values ​​respectively, and determining the surface where the first geometric points corresponding to the first distance values ​​with the largest number are located as the reference surface of the three-dimensional wall model;

[0073] The reference plane is marked in the three-dimensional model of the wall, and the virtual grinding and scraping robot is controlled to perform grinding and scraping processing on the grinding area based on the reference plane, and the operating parameter information of the grinding and scraping robot is collected in real time during the grinding and scraping processing of the grinding area by the virtual grinding and scraping robot to generate the initial grinding and scraping scheme; specifically, the reference plane is marked with a color different from that of the three-dimensional model of the wall, and the virtual grinding and scraping robot is controlled to grind the grinding area to the reference plane during the grinding and scraping processing; wherein the operating parameter information includes the execution terminal type of the grinding and scraping robot, the position of the execution terminal in the grinding area, the grinding angle of the execution terminal, the force applied by the execution terminal, etc. It should be noted that the grinding and scraping robot includes a plurality of execution terminals, and the shapes and sizes of the execution terminals are different from each other. During the grinding and scraping processing of the grinding area, the grinding and scraping robot automatically replaces the execution terminal according to the specific needs of grinding, and the grinding and scraping robot collects the powder generated by grinding while grinding.

[0074] It can be understood that the above-mentioned method of controlling the virtual grinding and scraping robot to perform grinding and scraping processing on the three-dimensional wall model, and generating the initial grinding and scraping plan in the process of the virtual grinding and scraping robot performing grinding and scraping processing on the three-dimensional wall model, by constructing a reference surface of the three-dimensional wall model directly in front of the three-dimensional wall model based on the external contour of the three-dimensional wall model, and generating the initial grinding and scraping plan based on the reference surface, helps to realize automatic grinding of walls with complex contours and improve the intelligence level of wall grinding.

[0075] Step S2: determine whether the initial grinding and scraping scheme can effectively grind and scrape the wall surface to be processed.

[0076] Wherein, the step S2 comprises the following steps:

[0077] Acquire second point cloud data information of the polished area after polishing and scraping, and acquire second distance values ​​between each second geometric point of the second point cloud data information and the reference surface; specifically, acquire the second point cloud data information by a virtual laser scanner set at the origin position of the coordinate system in step S1;

[0078] Obtaining an initial standard deviation between each of the second distance values, and comparing the initial standard deviation with a preset standard deviation;

[0079] If the initial standard deviation is less than the preset standard deviation, it is determined that the initial grinding and scraping scheme can effectively grind and scrape the wall surface to be processed;

[0080] If the initial standard deviation is not less than the preset standard deviation, it is determined that the initial grinding and scraping scheme cannot effectively grind and scrape the wall surface to be processed.

[0081] It can be understood that by judging whether the initial grinding and scraping scheme is capable of effectively grinding and scraping the wall surface to be treated, it is helpful to achieve fine grinding of the wall surface to be treated.

[0082] Step S3: If not, perform reinforcement learning on the initial grinding and scraping scheme based on a preset reinforcement learning algorithm to obtain a target grinding and scraping scheme;

[0083] The method of performing reinforcement learning on the initial grinding and scraping scheme based on a preset reinforcement learning algorithm to obtain a target grinding and scraping scheme includes the following steps:

[0084] The three-dimensional wall model after the grinding and scraping process is input into a preset wall defect detection model to obtain defect information of the three-dimensional wall model after the grinding and scraping process; wherein the wall defect detection model is a pre-trained neural network model, and the defect information is the defects marked in the three-dimensional wall model after the wall defect detection model detects the three-dimensional wall model, and the types of each defect, wherein the defect types include protrusions and depressions;

[0085] Generate a repair plan for the three-dimensional wall model after grinding and scraping based on the defect information; specifically, for each defect, if the defect is a protrusion, the repair plan for the defect is to continue grinding the defect, and if the defect is a depression, the repair plan for the defect is to fill the defect;

[0086] The initial grinding and scraping scheme is optimized based on the repair scheme to obtain an intermediate grinding and scraping scheme; specifically, for each of the defects, the position of the defect is determined in the three-dimensional wall model; if the repair scheme for the defect is to continue grinding the defect, the virtual grinding and scraping robot is controlled to grind the defect based on the reference surface; if the repair scheme for the defect is to fill the defect, the virtual grinding and scraping robot is controlled to fill the defect based on the reference surface, the position of the defect is recorded, and the repair operation parameter information of the virtual grinding and scraping robot when repairing the defect is collected, and the correlation between the position of the defect and the repair operation parameter information is generated, and the initial grinding and scraping scheme is optimized based on the correlation to obtain an intermediate grinding and scraping scheme;

[0087] Controlling the virtual grinding and scraping robot to perform grinding and scraping processing on the grinding area based on the intermediate grinding and scraping scheme, and obtaining third point cloud data information of the grinding area after the grinding and scraping processing, and obtaining third distance values ​​between each third geometric point of the third point cloud data information and the reference plane;

[0088] Obtaining an intermediate standard deviation between each of the third distance values, and comparing the intermediate standard deviation with a preset standard deviation;

[0089] If the intermediate standard deviation is not less than the preset standard deviation, the three-dimensional wall model after grinding and scraping is input into the preset wall defect detection model in an iterative loop to obtain the defect information of the three-dimensional wall model after grinding and scraping and subsequent steps until the intermediate standard deviation is less than the preset standard deviation to obtain the target grinding and scraping plan.

[0090] The training method of the wall defect detection model comprises the following steps:

[0091] Acquire a training wall three-dimensional model and label information corresponding to the training wall three-dimensional model; the label information is defects marked on the training wall three-dimensional model and the type of each defect;

[0092] The training wall three-dimensional model and the label information are used as a training pair and input into a neural network model for training;

[0093] Obtaining a test wall three-dimensional model and standard label information corresponding to the test wall three-dimensional model;

[0094] Inputting the test wall three-dimensional model into the trained neural network model to obtain predicted label information of the test wall three-dimensional model;

[0095] Obtaining a loss value between the standard label information and the predicted label information;

[0096] comparing the loss value with a preset loss value;

[0097] If the loss value is greater than the preset loss value, reversely adjusting the model parameters of the trained neural network model based on the loss value;

[0098] The above training method is iterated until the loss value is no greater than the preset loss value, and the trained neural network model is used as the wall defect detection model.

[0099] It can be understood that when the neural network model is iteratively trained, the training wall three-dimensional models corresponding to each training are different.

[0100] Step S4: Based on the target grinding and scraping scheme, control the grinding and scraping robot to grind and scrape the wall surface to be processed.

[0101] The method provided in this embodiment can, on the one hand, realize the grinding and scraping of complex wall surfaces and improve the intelligence level of wall grinding and scraping; on the other hand, it can realize the refined grinding and scraping processing of wall surfaces and improve the aesthetics of the wall surfaces.

[0102] See also Figure 2 , Figure 2 A schematic block diagram of the structure of a grinding and scraping robot 100 based on a learning algorithm provided in an embodiment of the present application is shown in FIG. Figure 2 As shown, the grinding and scraping robot 100 based on the learning algorithm provided in the embodiment of the present application includes:

[0103] The first generation module 110 is used to construct a three-dimensional wall model of the wall to be processed, and generate an initial grinding and scraping scheme of the three-dimensional wall model based on a preset initial grinding and scraping scheme generation method;

[0104] A judgment module 120 is used to judge whether the initial grinding and scraping scheme can effectively grind and scrape the wall surface to be processed;

[0105] The second generation module 130 is used to perform reinforcement learning on the initial grinding and scraping scheme based on a preset reinforcement learning algorithm to obtain a target grinding and scraping scheme if the initial grinding and scraping scheme cannot effectively grind and scrape the wall surface to be processed;

[0106] The control module 140 is used to control the grinding and scraping robot to grind and scrape the wall surface to be processed based on the target grinding and scraping scheme.

[0107] It should be noted that technical personnel in the relevant technical field can clearly understand that, for the convenience and conciseness of description, the specific working process of the robot and each module described above can refer to the corresponding process in the aforementioned embodiment of the grinding and scraping robot control method based on learning algorithm, and will not be repeated here.

[0108] The grinding and scraping robot 100 based on the learning algorithm provided in the above embodiment can be implemented in the form of a computer program. The computer program can be used in Figure 3 The system is run on the terminal device 200 shown.

[0109] See also Figure 3 , Figure 3 The present invention provides a schematic block diagram of the structure of a terminal device 200 according to an embodiment of the present application. The terminal device 200 includes a processor 201 and a memory 202. The processor 201 and the memory 202 are connected via a device bus 203, wherein the memory 202 may include a non-volatile storage medium and an internal memory.

[0110] The non-volatile storage medium can store a computer program. The computer program includes program instructions, and when the program instructions are executed by the processor 201, the processor 201 can execute any of the above-mentioned methods for controlling a grinding and scraping robot based on a learning algorithm.

[0111] The processor 201 is used to provide computing and control capabilities to support the operation of the entire terminal device 200.

[0112] The internal memory provides an environment for the operation of the computer program in the non-volatile storage medium. When the computer program is executed by the processor 201, the processor 201 can execute any of the above-mentioned grinding and scraping robot control methods based on the learning algorithm.

[0113] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a partial structure related to the scheme of the present application, and does not constitute a limitation on the terminal device 200 involved in the scheme of the present application. The specific terminal device 200 may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0114] It should be understood that the processor 201 may be a central processing unit (CPU), and the processor 201 may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0115] In some embodiments, the processor 201 is used to run a computer program stored in the memory to implement the following steps:

[0116] Constructing a three-dimensional wall model of the wall to be processed, and generating an initial grinding and scraping scheme of the three-dimensional wall model based on a preset initial grinding and scraping scheme generation method;

[0117] Determine whether the initial grinding and scraping scheme can effectively grind and scrape the wall surface to be processed;

[0118] If not, the initial grinding and scraping scheme is subjected to reinforcement learning based on a preset reinforcement learning algorithm to obtain a target grinding and scraping scheme;

[0119] Based on the target grinding and scraping scheme, the grinding and scraping robot is controlled to grind and scrape the wall surface to be processed.

[0120] It should be noted that technical personnel in the relevant field can clearly understand that for the convenience and simplicity of description, the specific working process of the terminal device 200 described above can refer to the corresponding process of the aforementioned grinding and scraping robot control method based on learning algorithm, and will not be repeated here.

[0121] An embodiment of the present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by one or more processors, the one or more processors implement the grinding and scraping robot control method based on the learning algorithm as provided in the embodiment of the present application.

[0122] The computer-readable storage medium may be an internal storage unit of the terminal device 200 in the aforementioned embodiment, such as a hard disk or memory of the terminal device 200. The computer-readable storage medium may also be an external storage device of the terminal device 200, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped with the terminal device 200.

[0123] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in the present application, and these modifications or replacements should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A control method for a grinding and scraping robot based on a learning algorithm, characterized in that: include: Constructing a three-dimensional wall model of the wall to be processed, and generating an initial grinding and scraping scheme of the three-dimensional wall model based on a preset initial grinding and scraping scheme generation method; Determine whether the initial grinding and scraping scheme can effectively grind and scrape the wall surface to be processed; If not, the initial grinding and scraping scheme is subjected to reinforcement learning based on a preset reinforcement learning algorithm to obtain a target grinding and scraping scheme; Controlling the grinding and scraping robot to grind and scrape the wall surface to be processed based on the target grinding and scraping scheme; The step of constructing a three-dimensional wall model of the wall to be processed and generating an initial grinding and scraping scheme of the three-dimensional wall model based on a preset initial grinding and scraping scheme generation method includes: Acquire image information and first point cloud data information of a wall surface to be processed, and generate a three-dimensional wall model of the wall surface to be processed based on the image information and the first point cloud data information; Controlling the virtual grinding and scraping robot to perform grinding and scraping processing on the three-dimensional wall model, and generating the initial grinding and scraping scheme in the process of the virtual grinding and scraping robot performing grinding and scraping processing on the three-dimensional wall model; wherein the virtual grinding and scraping robot is matched with the grinding and scraping robot; The controlling the virtual grinding and scraping robot to perform grinding and scraping processing on the three-dimensional wall model, and generating the initial grinding and scraping scheme in the process of the virtual grinding and scraping robot performing grinding and scraping processing on the three-dimensional wall model, includes: Determining a polishing area of ​​the three-dimensional wall model; Constructing a reference surface of the three-dimensional wall model directly in front of the three-dimensional wall model based on the outer contour of the three-dimensional wall model; Obtaining first distance values ​​between each first geometric point in the polishing area and the reference surface; Counting the number of each of the first distance values ​​respectively, and determining the surface where the first geometric points corresponding to the first distance values ​​with the largest number are located as the reference surface of the three-dimensional wall model; The reference plane is marked in the three-dimensional model of the wall, and the virtual grinding and scraping robot is controlled to perform grinding and scraping on the grinding area based on the reference plane. In the process of the virtual grinding and scraping robot performing grinding and scraping on the grinding area, the operating parameter information of the grinding and scraping robot is collected in real time to generate the initial grinding and scraping plan.

2. The control method of a grinding and scraping robot based on a learning algorithm according to claim 1 is characterized in that: The step of acquiring image information and first point cloud data information of a wall surface to be processed, and generating a three-dimensional wall model of the wall surface to be processed based on the image information and the first point cloud data information, includes: The image information is acquired by a camera device preset on the polishing and scraping robot, and the first point cloud data information is acquired by a laser scanner preset on the polishing and scraping robot; The geometric structure of the wall to be processed is constructed based on the first point cloud data information, and the geometric structure is rendered based on the image information to obtain the three-dimensional model of the wall.

3. The control method of a grinding and scraping robot based on a learning algorithm according to claim 1, characterized in that: The determining whether the initial grinding and scraping scheme can effectively grind and scrape the wall surface to be processed includes: Acquire second point cloud data information of the polished area after polishing and scraping, and acquire second distance values ​​between each second geometric point of the second point cloud data information and the reference surface; Obtaining an initial standard deviation between each of the second distance values, and comparing the initial standard deviation with a preset standard deviation; If the initial standard deviation is less than the preset standard deviation, it is determined that the initial grinding and scraping scheme can effectively grind and scrape the wall surface to be processed; If the initial standard deviation is not less than the preset standard deviation, it is determined that the initial grinding and scraping scheme cannot effectively grind and scrape the wall surface to be processed.

4. The control method of the grinding and scraping robot based on the learning algorithm according to claim 3 is characterized in that: The initial grinding and scraping scheme is subjected to reinforcement learning based on a preset reinforcement learning algorithm to obtain a target grinding and scraping scheme, including: Inputting the three-dimensional wall model after the grinding and scraping process into a preset wall defect detection model to obtain defect information of the three-dimensional wall model after the grinding and scraping process; wherein the wall defect detection model is a pre-trained neural network model; Generating a repair plan for the three-dimensional model of the wall after grinding and scraping based on the defect information; Optimizing the initial grinding and scraping scheme based on the repair scheme to obtain an intermediate grinding and scraping scheme; Controlling the virtual grinding and scraping robot to perform grinding and scraping processing on the grinding area based on the intermediate grinding and scraping scheme, and obtaining third point cloud data information of the grinding area after the grinding and scraping processing, and obtaining third distance values ​​between each third geometric point of the third point cloud data information and the reference plane; Obtaining an intermediate standard deviation between each of the third distance values, and comparing the intermediate standard deviation with a preset standard deviation; If the intermediate standard deviation is not less than the preset standard deviation, the three-dimensional wall model after grinding and scraping is input into the preset wall defect detection model in an iterative loop to obtain the defect information of the three-dimensional wall model after grinding and scraping and subsequent steps until the intermediate standard deviation is less than the preset standard deviation to obtain the target grinding and scraping plan.

5. The control method of a grinding and scraping robot based on a learning algorithm according to claim 4 is characterized in that: The training method of the wall defect detection model includes: Obtaining a training wall three-dimensional model and label information corresponding to the training wall three-dimensional model; The training wall three-dimensional model and the label information are used as a training pair and input into a neural network model for training; Obtaining a test wall three-dimensional model and standard label information corresponding to the test wall three-dimensional model; Inputting the test wall three-dimensional model into the trained neural network model to obtain predicted label information of the test wall three-dimensional model; Obtaining a loss value between the standard label information and the predicted label information; comparing the loss value with a preset loss value; If the loss value is greater than the preset loss value, reversely adjusting the model parameters of the trained neural network model based on the loss value; The above training method is iterated until the loss value is no greater than the preset loss value, and the trained neural network model is used as the wall defect detection model.

6. A grinding and scraping robot based on a learning algorithm, characterized in that: include: A first generation module is used to construct a three-dimensional wall model of the wall to be processed, and generate an initial grinding and scraping scheme of the three-dimensional wall model based on a preset initial grinding and scraping scheme generation method; A judgment module, used to judge whether the initial grinding and scraping scheme can effectively grind and scrape the wall surface to be processed; The second generation module is used for performing reinforcement learning on the initial grinding and scraping scheme based on a preset reinforcement learning algorithm to obtain a target grinding and scraping scheme if the initial grinding and scraping scheme cannot effectively grind and scrape the wall surface to be treated; A control module, used for controlling the grinding and scraping robot to grind and scrape the wall surface to be processed based on the target grinding and scraping scheme; The step of constructing a three-dimensional wall model of the wall to be processed and generating an initial grinding and scraping scheme of the three-dimensional wall model based on a preset initial grinding and scraping scheme generation method includes: Acquire image information and first point cloud data information of a wall surface to be processed, and generate a three-dimensional wall model of the wall surface to be processed based on the image information and the first point cloud data information; Controlling the virtual grinding and scraping robot to perform grinding and scraping processing on the three-dimensional wall model, and generating the initial grinding and scraping scheme in the process of the virtual grinding and scraping robot performing grinding and scraping processing on the three-dimensional wall model; wherein the virtual grinding and scraping robot is matched with the grinding and scraping robot; The controlling the virtual grinding and scraping robot to perform grinding and scraping processing on the three-dimensional wall model, and generating the initial grinding and scraping scheme in the process of the virtual grinding and scraping robot performing grinding and scraping processing on the three-dimensional wall model, includes: Determining a polishing area of ​​the three-dimensional wall model; Constructing a reference surface of the three-dimensional wall model directly in front of the three-dimensional wall model based on the outer contour of the three-dimensional wall model; Obtaining first distance values ​​between each first geometric point in the polishing area and the reference surface; Counting the number of each of the first distance values ​​respectively, and determining the surface where the first geometric points corresponding to the first distance values ​​with the largest number are located as the reference surface of the three-dimensional wall model; The reference plane is marked in the three-dimensional model of the wall, and the virtual grinding and scraping robot is controlled to perform grinding and scraping on the grinding area based on the reference plane. In the process of the virtual grinding and scraping robot performing grinding and scraping on the grinding area, the operating parameter information of the grinding and scraping robot is collected in real time to generate the initial grinding and scraping plan.

Images