Control method, device and equipment of guniting robot and medium

Abstract

The invention discloses a control method, device and equipment of a guniting robot and a medium. The residual path length of a target roadway is determined through a guniting robot so as to determine the guniting operation progress of the target roadway, and a plurality of side wall sub-areas needing guniting are determined based on an initial roadway side wall image collected at each target position point in an image recognition mode so as to generate a plurality of area spraying control parameters for spraying control. And identifying a guniting completion state based on the acquired target roadway side wall image, and when guniting on the roadway side wall area of the target position point is completed, continuing to advance to the next target position point to execute guniting operation until the guniting operation is completed, so that a manual guniting mode or a mode of manually controlling the remote control robot is not needed, and the guniting efficiency is improved. The self-adaptive guniting operation is carried out on the target roadway through the guniting robot, flexibility and convenience are achieved, the guniting operation efficiency of the target roadway is improved, the operation duration is shortened, and the guniting operation progress of the target roadway is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computers, and in particular to a control method and device for a shotcrete robot, an equipment and a medium. BACKGROUND

[0002] In the field of coal mine industry, in order to provide a safe and stable working environment, it is necessary to reinforce the mine roadway, such as using concrete, building mortar and other civil engineering materials to coat the walls of the mine roadway, so as to reinforce the walls of the mine roadway and provide a safe and reliable coal mine site working environment.

[0003] The related art generally uses manual coating to coat the mine roadway, or uses a remote control robot to coat the walls of the mine roadway, so that the mine roadway can be reinforced to provide a safe and reliable coal mine site working environment.

[0004] However, the related art has low efficiency by using manual coating, and long-term work in high dust conditions can endanger the personal safety of workers, and using a remote control robot to replace manual work lacks flexibility and is also low in efficiency, which affects the coating progress. SUMMARY

[0005] The embodiments of the present application provide a control method and device for a shotcrete robot, an equipment and a medium, which can automatically adapt to coating work without human control, have flexibility, improve coating efficiency, and thus improve coating progress.

[0006] Improve the accuracy of visual target detection of an unmanned boat in a wave fluctuation environment.

[0007] In a first aspect, the present application provides a control method for a shotcrete robot, applied to a shotcrete robot including a shotcrete assembly and a motor assembly, comprising: determining a remaining path length corresponding to a target roadway; when the remaining path length is greater than a preset length threshold, controlling the motor assembly to travel to a next target position point in the target roadway, and collecting an initial roadway side wall image of the surrounding roadway side wall of the target position point; determining a plurality of side wall sub-regions corresponding to the target position point according to the initial roadway side wall image, and generating a plurality of regional spraying control parameters according to the plurality of side wall sub-regions; controlling the shotcrete assembly to sequentially spray the plurality of side wall sub-regions according to the plurality of regional spraying control parameters, and collecting a target roadway side wall image of the surrounding roadway side wall of the target position point after spraying; Based on the target tunnel sidewall image, determine the shotcreting completion status of the multiple sidewall sub-regions; When the shotcreting completion status is "shotcreting complete", return to the step of determining the remaining path length corresponding to the target tunnel, and stop the shotcreting operation on the target tunnel when the corresponding remaining path length is less than or equal to the preset length threshold.

[0008] Secondly, this application provides a control device for a shotcrete robot, applied to a shotcrete robot, the shotcrete robot including a shotcrete assembly and a motor assembly, comprising: The determination unit is used to determine the remaining path length corresponding to the target roadway; The first control unit is used to control the motorized component to travel to the next target location point in the target tunnel when the remaining path length is greater than a preset length threshold, and to acquire the initial tunnel sidewall image of the surrounding tunnel sidewall of the target location point. The generation unit is used to determine multiple sidewall sub-regions corresponding to the target location point based on the initial tunnel sidewall image, and to generate multiple area spraying control parameters based on the multiple sidewall sub-regions. The second control unit controls the spraying assembly to spray the multiple sidewall sub-regions sequentially according to the multiple area spraying control parameters, and acquires images of the surrounding roadway sidewalls of the target location point after spraying. The status detection unit is used to determine the shotcreting completion status of the multiple sidewall sub-regions based on the target tunnel sidewall image. The repeat execution unit is used to return to the step of determining the remaining path length corresponding to the target tunnel when the shotcreting completion status is shotcreting complete, until the corresponding remaining path length is less than or equal to the preset length threshold and then stop the shotcreting operation on the target tunnel.

[0009] In some embodiments, the generating unit is further configured to: The target tunnel sidewall region is determined based on the initial tunnel sidewall image, and the target total area corresponding to the target tunnel sidewall region is determined. From the list of preset division coefficients, query the target division coefficient corresponding to the target total area. The list of preset division coefficients includes multiple different preset total areas and the preset division coefficient corresponding to each preset total area. The target roadway sidewall region is divided according to the target division coefficient to obtain multiple sidewall sub-regions.

[0010] In some embodiments, the state detection unit is further configured to: The target tunnel sidewall image is identified using a pre-trained target image recognition model to obtain the recognition result. The pre-trained target image recognition model is obtained by minimizing the difference between the predicted recognition result output by the sample target alleyway sidewall image and the sample recognition result. Based on the identification results, the shotcreting completion status corresponding to the multiple sidewall sub-regions is determined.

[0011] In some embodiments, the control device of the shotcrete robot further includes a third control unit for: Identify at least one target similarity region between the target tunnel sidewall image and the initial tunnel sidewall image; Based on the target coordinate information of the target similar area, generate corresponding target area spraying control parameters; According to the target area spraying control parameters, the spraying assembly is controlled to spray and repair the corresponding target-similar area on the sidewall surrounding the target location point; Collect images of the sidewalls of the surrounding roadway at the target location after shotcrete repair; Based on the image of the sidewall to be identified, the shotcrete completion status of the multiple sidewall sub-regions is reassessed.

[0012] In some embodiments, the third control unit is further configured to: The image of the side wall to be identified is input into a pre-trained target image recognition model for recognition, and a first recognition result is obtained. Based on the first recognition result, a first similarity is determined between the sidewall image to be recognized and the initial tunnel sidewall image; When the first similarity is greater than a preset similarity threshold, a second similarity is determined between the sidewall image to be identified and the target alleyway sidewall image. When the second similarity is less than or equal to the preset spraying similarity threshold, the spraying completion status corresponding to the plurality of sidewall sub-regions is determined to be spraying completion; When the second similarity is greater than the preset spraying similarity threshold, the spraying completion status of the multiple sidewall sub-regions is determined to be spraying incomplete.

[0013] In some embodiments, the generating unit is further configured to: Based on the initial tunnel sidewall image, identify the surface smoothness of the corresponding area wall for each sidewall sub-region; Based on the smoothness of the wall surface in each area, determine the spraying pressure parameters for each sidewall sub-area; Determine the region coordinate information corresponding to each sidewall sub-region; By combining the region coordinate information and the spraying pressure parameters corresponding to each sidewall sub-region, multiple region spraying control parameters are generated.

[0014] In some embodiments, the state detection unit is further configured to: The target tunnel sidewall image is input into a pre-trained target image recognition model for recognition, and the target recognition result is obtained. The point cloud information of each side wall sub-region is obtained by LiDAR, and the smoothness of the wall surface of each side wall sub-region is determined based on the point cloud information of each region. When it is determined that spraying is not required based on the target recognition result, and the smoothness of the wall surface in the area is greater than or equal to a preset smoothness threshold, the spraying completion status corresponding to the multiple side wall sub-areas is determined to be spraying completion. When it is determined that spraying is required based on the target recognition result, or when the smoothness of the wall surface in the area is less than the preset smoothness threshold, the spraying completion status of the multiple side wall sub-areas is determined to be spraying incomplete.

[0015] Furthermore, this application also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the above-described control method for the shotcrete robot.

[0016] Furthermore, embodiments of this application also provide a computer-readable storage medium storing a plurality of instructions adapted for loading by a processor to execute the aforementioned control method for a shotcrete robot.

[0017] In this embodiment of the application, when the shotcrete robot performs shotcrete operation on a target tunnel, it first determines the remaining path length corresponding to the target tunnel. When the remaining path length is greater than a preset length threshold, it controls the motorized component of the shotcrete robot to travel to the next target location point within the target tunnel to acquire an initial tunnel sidewall image of the surrounding tunnel sidewalls at the target location point. Then, based on the initial tunnel sidewall image, it determines multiple sidewall sub-regions corresponding to the target location point, and generates multiple area spraying control parameters for these multiple sidewall sub-regions. Next, according to these multiple area spraying control parameters, it controls the shotcrete robot's shotcrete component to spray onto the multiple sidewalls. Shotcrete operations are performed in sub-regions to acquire images of the target roadway sidewalls surrounding the target location after shotcrete. Finally, based on the acquired target roadway sidewall images, the shotcrete completion status of these multiple sidewall sub-regions is determined. If the shotcrete completion status is "shotcrete complete", the process returns to the step of determining the remaining path length of the target roadway and continues shotcrete operations on the sidewalls surrounding the next target location of the target roadway until the remaining path length is less than or equal to a preset length threshold. This is considered as completion of shotcrete operations on all roadway sidewalls of the target roadway. In this way, automatic shotcrete operations on the target roadway are achieved.

[0018] Therefore, compared to the inefficient manual grouting method in related technologies, which is not only inefficient but also poses a safety hazard to workers under high dust conditions, and the inefficient method of using remote-controlled robots to replace manual labor, which requires human operation and thus lacks flexibility and affects the progress of grouting operations, this application uses a spraying robot to determine the remaining path length of the target tunnel, thereby determining the progress of the spraying operation. Through image recognition, based on the initial tunnel sidewall images collected at each target location, multiple sidewall sub-regions requiring spraying are identified, generating multiple area spraying control parameters for spraying control. The spraying completion status is identified based on the collected target tunnel sidewall images. When the spraying of the tunnel sidewall area at a target location is completed, the robot continues to the next target location to perform the spraying operation until completion. Thus, there is no need for manual grouting or human operation of the remote-controlled robot; instead, the spraying robot performs adaptive spraying operations on the target tunnel, offering flexibility and convenience, improving the efficiency of the spraying operation, reducing operation time, and increasing the progress of the spraying operation. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 A schematic diagram of a control system for a shotcrete robot provided in an embodiment of this application; Figure 2 A schematic diagram of the hardware structure of the shotcrete robot provided in the embodiments of this application; Figure 3 This is a flowchart illustrating the steps of the control method for the shotcrete robot provided in the embodiments of this application; Figure 4 Example diagram of the target alleyway scene provided in the embodiments of this application; Figure 5 Example diagram of sidewall sub-region division scenario provided in the embodiments of this application; Figure 6 This is an example diagram of the control flow of the shotcrete robot provided in the embodiments of this application; Figure 7 This is a schematic diagram of the control device for the shotcrete robot provided in the embodiments of this application; Figure 8 This is a schematic diagram of the terminal structure provided in an embodiment of this application. Detailed Implementation

[0021] To enable those skilled in the art to better understand the solutions of this application, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0022] It should be noted that while some processes described in the specification, claims, and accompanying drawings contain multiple steps that appear in a specific order, it should be clearly understood that these steps may not be performed in the order they appear herein, or may be performed in parallel. The step numbers are merely used to distinguish different steps and do not represent any particular order of execution. Furthermore, descriptions such as "first," "second," or "objective" in this document are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0023] It is understood that in the specific embodiments of this application, data such as the remaining path length, initial tunnel sidewall image, area spraying control parameters, target tunnel sidewall image, and spraying completion status are involved. When the above embodiments of this application are applied to specific products or technologies, permission or consent from the target is required, and the collection, use and processing of related data must comply with relevant laws, regulations and standards.

[0024] Furthermore, when this application embodiment needs to obtain relevant data, it will obtain separate permission or separate consent for data such as the remaining path length, initial tunnel sidewall image, area spraying control parameters, target tunnel sidewall image, and spraying completion status through pop-up windows or redirection to a confirmation page. Only after clearly obtaining separate permission or separate consent for the relevant data such as the remaining path length, initial tunnel sidewall image, area spraying control parameters, target tunnel sidewall image, and spraying completion status will the necessary data for enabling the application embodiment to operate normally be obtained.

[0025] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0026] This application provides a control method, apparatus, device, and medium for a shotcrete robot. Specifically, the control method for the shotcrete robot in this application can be implemented in a computer device, which can be a terminal device. The terminal device can be a robot, aircraft, drone, etc., but is not limited to these.

[0027] For ease of understanding, this application will describe the implementation process of the control method for the shotcrete robot through several embodiments, as follows: This application provides a control method for a shotcrete robot. The shotcrete robot determines the remaining path length of a target tunnel to ascertain the shotcrete operation progress. Using image recognition, it identifies multiple sub-regions of the tunnel sidewall requiring shotcreting based on initial images of the tunnel sidewall at each target location. This generates multiple area spraying control parameters for control. The method also identifies the shotcrete completion status based on the acquired images of the target tunnel sidewall. When shotcreting of the tunnel sidewall region at a target location is complete, the robot continues to the next target location to perform shotcreting operations until completion. Please refer to the following specific embodiments for details.

[0028] For example, see Figure 1 This is a schematic diagram of a control system for a shotcrete robot provided in an embodiment of this application. The system includes a terminal 110.

[0029] Each terminal 110 may have a target application installed on it, and the corresponding application business can be run through the target application. The target application may be called a client, and the terminal 110 may be a shotcrete robot, an aircraft, or a drone. For example, taking terminal 110 as a shotcrete robot, the shotcrete robot can determine the remaining path length corresponding to the target tunnel; when the remaining path length is greater than a preset length threshold, it controls the motorized component to travel to the next target location point in the target tunnel and collects the initial tunnel sidewall image of the surrounding tunnel sidewall of the target location point; based on the initial tunnel sidewall image, it determines multiple sidewall sub-regions corresponding to the target location point and generates multiple area spraying control parameters based on the multiple sidewall sub-regions; based on the multiple area spraying control parameters, it controls the shotcrete component to sequentially spray the multiple sidewall sub-regions and collects the target tunnel sidewall image of the surrounding tunnel sidewall of the target location point after spraying; based on the target tunnel sidewall image, it determines the shotcrete completion status corresponding to the multiple sidewall sub-regions; when the shotcrete completion status is shotcrete complete, it returns to determine the remaining path length corresponding to the target tunnel until the corresponding remaining path length is less than or equal to the preset length threshold, at which point the shotcrete operation on the target tunnel stops.

[0030] This application also provides a shotcrete robot. Figure 2 This is a schematic diagram of the hardware structure of the shotcrete robot provided in an embodiment of this application. (In conjunction with...) Figure 2 As shown, the spraying robot includes a camera assembly, a motor assembly, and a spraying assembly. The camera assembly may include multiple cameras, such as... Figure 2 As shown, the shotcrete robot includes four cameras; the motorized component includes multiple pulleys, such as four pulleys; the spraying component includes a nozzle, a spray bar angle control joint, and a shotcrete interface; in addition, the shotcrete robot also includes a navigation and detection controller, which is used to execute the control method of the shotcrete robot.

[0031] In accordance with the above method, when the shotcrete robot proposed in this embodiment performs shotcrete operation on a target tunnel, it first determines the remaining path length corresponding to the target tunnel. When the remaining path length is greater than a preset length threshold, it controls the motorized component of the shotcrete robot to travel to the next target location point within the target tunnel to acquire initial tunnel sidewall images of the surrounding tunnel sidewalls at the target location point. Then, based on the initial tunnel sidewall images, it determines multiple sidewall sub-regions corresponding to the target location point, and generates multiple area spraying control parameters for these multiple sidewall sub-regions. Finally, it controls the shotcrete component of the shotcrete robot according to these multiple area spraying control parameters. Shotcrete operations are performed on multiple sidewall sub-regions to acquire images of the target roadway sidewalls surrounding the target location after shotcrete. Finally, based on the acquired target roadway sidewall images, the shotcrete completion status corresponding to these multiple sidewall sub-regions is determined. If the shotcrete completion status is "shotcrete complete", the process returns to the step of determining the remaining path length corresponding to the target roadway, and shotcrete operations are continued on the sidewalls surrounding the next target location in the target roadway until the remaining path length is less than or equal to a preset length threshold. This is considered as completion of shotcrete operations on all roadway sidewalls of the target roadway. In this way, automatic shotcrete operations on the target roadway are achieved.

[0032] Therefore, compared to the inefficient manual grouting method in related technologies, which is not only inefficient but also poses a safety hazard to workers under high dust conditions, and the inefficient method of using remote-controlled robots to replace manual labor, which requires human operation and thus lacks flexibility and affects the progress of grouting operations, this application uses a spraying robot to determine the remaining path length of the target tunnel, thereby determining the progress of the spraying operation. Through image recognition, based on the initial tunnel sidewall images collected at each target location, multiple sidewall sub-regions requiring spraying are identified, generating multiple area spraying control parameters for spraying control. The spraying completion status is identified based on the collected target tunnel sidewall images. When the spraying of the tunnel sidewall area at a target location is completed, the robot continues to the next target location to perform the spraying operation until completion. Thus, there is no need for manual grouting or human operation of the remote-controlled robot; instead, the spraying robot performs adaptive spraying operations on the target tunnel, offering flexibility and convenience, improving the efficiency of the spraying operation, reducing operation time, and increasing the progress of the spraying operation.

[0033] For ease of understanding, the steps of the control method for the shotcrete robot will be described in detail below. It should be noted that the order of the following embodiments is not intended to limit the preferred order of the embodiments.

[0034] See Figure 3 , Figure 3This is a flowchart illustrating the steps of a control method for a shotcrete robot provided in an embodiment of this application. In this embodiment, the control method for the shotcrete robot can be executed by a computer device, such as a terminal. Specifically, taking the shotcrete robot as an example, the control method for the shotcrete robot is executed as follows: 101. Determine the remaining path length corresponding to the target roadway.

[0035] In the coal mining industry, to provide a safe and stable working environment, it is necessary to use civil engineering materials such as concrete and construction mortar to coat the walls of mine roadways. This reinforces the walls, prevents collapse, and thus provides a safe and reliable working environment in coal mines. Related technologies generally employ manual coating methods or remote-controlled robots to coat the walls of mine roadways. However, manual coating is not only inefficient but also poses a safety hazard to workers under prolonged high-dust conditions. While using remote-controlled robots requires human operation, it lacks flexibility and is also inefficient, affecting the progress of the coating operation.

[0036] To address the above issues, this application embodiment uses a shotcrete robot to determine the remaining path length of the target tunnel, thereby determining the shotcrete operation progress of the target tunnel. Through image recognition, based on the initial tunnel sidewall images collected at each target location, multiple sidewall sub-regions requiring shotcrete are identified to generate multiple area spraying control parameters for spraying control. The shotcrete completion status is also identified based on the collected target tunnel sidewall images. When the shotcrete of the tunnel sidewall region at the target location is completed, the robot continues to the next target location to perform shotcrete operations until completion. Thus, there is no need for manual grouting or manual control of the remote-controlled robot; instead, the shotcrete robot performs adaptive shotcrete operations on the target tunnel, offering flexibility and convenience, improving the efficiency of shotcrete operations on the target tunnel, reducing operation time, and increasing the shotcrete operation progress of the target tunnel.

[0037] The shotcrete robot can automatically determine the remaining path length of the target roadway in the current coal mine environment to assess the work progress. It also uses a camera module to capture initial and post-shotcrete images of the roadway sidewalls surrounding each target location to determine the completion status of the shotcrete application. This automated shotcrete operation on the target roadway sidewalls is achieved through this process, offering flexibility and improving shotcrete efficiency. To facilitate understanding of the shotcrete robot provided in this embodiment, combined with... Figure 2 As shown, the spraying robot includes a camera assembly, a motor assembly, and a spraying assembly. The camera assembly may include multiple cameras, such as... Figure 2As shown, the shotcrete robot includes four cameras; the motorized component includes multiple pulleys, such as four pulleys; the spraying component includes a nozzle, a spray bar angle control joint, and a shotcrete interface; in addition, the shotcrete robot includes an explosion-proof motor for driving the four pulleys to enable the shotcrete robot to move, and also includes a navigation and detection controller, which is used to execute the control method of the shotcrete robot.

[0038] The remaining path length can be the path length of the current target tunnel in the forward direction, which can be understood as the remaining tunnel length for which spraying operations need to be performed. It is understood that in the spraying operation of this embodiment, the spraying robot advances a unit distance in the target tunnel to reach a target location point, and then performs spraying operations on the tunnel sidewalls surrounding that target location. Therefore, the remaining path length can be understood as the path length for which the spraying robot needs to move forward.

[0039] In some implementations, the distance between the current position of the shotcrete robot in the target tunnel and the end of the target tunnel can be determined by lidar ranging as the remaining path length. For example, step 101 may include: determining the remaining path length corresponding to the target tunnel using lidar.

[0040] By using the above method, the remaining path length corresponding to the target tunnel can be determined, so that the remaining path length can be determined based on the remaining path length. The remaining path length can reflect the shotcreting progress of the target tunnel, so that the shotcreting robot can decide whether to continue the shotcreting operation based on the remaining path length.

[0041] 102. When the remaining path length is greater than the preset length threshold, control the motor component to travel to the next target location point in the target tunnel, and collect the initial tunnel sidewall image of the surrounding tunnel sidewall of the target location point.

[0042] In this embodiment, after obtaining the remaining path length, it can be compared with a preset length threshold to determine the shotcreting progress for the target tunnel based on the identification result. For example, if the remaining path length is greater than the preset length threshold, it is determined that the shotcreting progress for the target tunnel is incomplete, and shotcreting operations need to continue on the sidewalls of the surrounding tunnel along the remaining path length. Therefore, the shotcreting robot can use a motor to control the rotation of the motorized components and move forward to the next target location point within the target tunnel. Furthermore, after reaching the target location point, an initial image of the sidewalls of the surrounding tunnel at that target location point is acquired. This allows for subsequent identification of the area of ​​the sidewalls surrounding the current target location based on the initial sidewall image, preparing for the shotcreting operation.

[0043] The preset length threshold can be a judgment value based on the remaining distance of the target tunnel. By comparing the remaining path length with the preset length threshold, the shotcreting progress of the target tunnel is determined. The preset length threshold can be determined according to the actual situation. For example, it can be set to 0. When the remaining path length is greater than 0, it means that the shotcreting robot can continue to move forward, that is, there are still areas on the sidewall of the tunnel that have not been shotcreted. Therefore, the shotcreting robot needs to continue to move forward to perform the subsequent shotcreting.

[0044] The target location can be a point on the target tunnel. After advancing a target unit distance, the shotcrete robot can reach the next location where shotcrete work needs to be performed. The target unit distance can be 0.3 meters or any value within the range of 0.1 to 0.5 meters. The target unit distance can be determined based on the spraying pressure parameters. For example, the current spraying pressure parameters can be determined, and the corresponding target unit distance can be retrieved from the preset displacement distance list based on the spraying pressure parameters. The preset displacement distance list contains multiple spraying pressure parameters and the preset unit distance corresponding to each spraying pressure parameter. Alternatively, it can be determined by combining the spraying pressure parameters and the shotcrete flow rate. In this way, the motorized component is controlled to travel to the next target location in the target tunnel based on the target distance parameters. This achieves the goal of determining the displacement of the shotcrete robot by combining the spraying pressure parameters and / or the shotcrete flow rate, so as to avoid the shotcrete robot's single displacement being too large, resulting in poor shotcrete effect on the tunnel sidewalls around the target location, such as uniform, smooth, and flat sidewall surfaces after spraying, and also avoids the shotcrete robot's single displacement being too small, which would affect the efficiency of the shotcrete operation, thus ensuring reliability.

[0045] The initial tunnel sidewall image can be an image of the tunnel sidewall region surrounding the current target location. It displays relevant visual information about the tunnel sidewalls around the current target location, such as their shape, curvature, and color. Based on this initial tunnel sidewall image, the basic condition of the tunnel sidewalls around the current target location can be understood, facilitating the subsequent identification of the shotcrete operation area.

[0046] Using the above method, the shotcreting progress for the target tunnel can be determined based on the comparison between the remaining path length and a preset length threshold. For example, if the remaining path length exceeds the preset length threshold, it is determined that shotcreting operations need to continue on the sidewalls of the tunnel surrounding the remaining path length of the target tunnel. Therefore, the shotcreting robot can be driven by a motor to rotate its motorized components, moving forward to the next target location within the target tunnel and acquiring initial images of the sidewalls surrounding that target location. This allows for subsequent identification of the area of ​​the sidewalls surrounding the current target location based on these initial sidewall images, preparing for the shotcreting operation.

[0047] 103. Determine multiple sidewall sub-regions corresponding to the target location points based on the initial roadway sidewall image, and generate multiple area spraying control parameters based on the multiple sidewall sub-regions.

[0048] In this embodiment, after obtaining the initial tunnel sidewall image, the target tunnel sidewall region at the target location point can be identified based on the initial tunnel sidewall image. The target tunnel sidewall region is then divided into multiple sub-regions for shotcreting operations. Corresponding spraying control parameters are generated for each sub-region, so that shotcreting operations can be performed on each sub-region based on the spraying control parameters. This shotcreting planning method, which divides the tunnel into multiple sub-regions, not only allows for adaptive shotcreting operations for different sub-regions (e.g., different shapes and curved surfaces can be targeted with different shotcreting operations, such as different shotcreting pressure parameters or different moving speeds of the shotcreting rod), but also facilitates subsequent shotcreting effect detection for each sub-region, enabling rapid location of areas requiring shotcreting repair and improving the efficiency of the entire shotcreting operation process.

[0049] Among them, multiple sidewall sub-regions refer to multiple sub-regions obtained by planning the target roadway sidewall region indicated by the initial roadway sidewall image. These multiple sidewall sub-regions are data information and do not actually physically decompose or split the target roadway sidewall region. Instead, the target roadway sidewall region is divided and represented by the processor of the shotcrete robot in the form of data. When dividing, it can be planned according to the equal division method or according to the area. There is no limitation here.

[0050] Among them, the multiple area spraying control parameters can be spraying control parameters corresponding to multiple side wall sub-areas. Specifically, each side wall sub-area corresponds to one area spraying control parameter. Each area spraying control parameter may include, but is not limited to, the position information of the corresponding side wall sub-area, spraying pressure parameters, and / or the moving speed of the spraying rod, etc. Each area spraying control parameter is used to instruct the spraying assembly to perform spraying operations on the corresponding side wall sub-area.

[0051] In some implementations, the corresponding target roadway sidewall region can be identified based on the initial roadway sidewall image, and the target roadway sidewall region can be divided into multiple sidewall sub-regions based on the target total area corresponding to the target roadway sidewall region. For example, step 103, "determining multiple sidewall sub-regions corresponding to the target location point based on the initial roadway sidewall image," can include: determining the corresponding target roadway sidewall region based on the initial roadway sidewall image, and determining the target total area corresponding to the target roadway sidewall region; querying the target division coefficient corresponding to the target total area from a preset division coefficient list; and dividing the target roadway sidewall region according to the target division coefficient to obtain multiple sidewall sub-regions.

[0052] The target tunnel sidewall region refers to the area of ​​the sidewall at the current target location identified based on the initial tunnel sidewall image, i.e., the area of ​​the tunnel sidewall surrounding the target location. This target tunnel sidewall region can be a 270-degree wall surface, combined with... Figure 4 The image shown is an example of a target tunnel scene provided in an embodiment of this application. When the spraying robot is at the target location point in the target tunnel, it identifies the target tunnel sidewall region around the target location point based on the initial tunnel sidewall image. This target tunnel sidewall region is an approximately annular wall at 270 degrees within the target tunnel, i.e., the spraying surface is the wall within the target tunnel. The above is only an example; other shapes are also possible, such as squares, rectangles, triangles, etc., and are not limited here.

[0053] The total target area refers to the spraying area corresponding to the side wall area of ​​the target tunnel at the target location point.

[0054] The target partitioning coefficient can be a numerical value used to indicate the division of the target roadway sidewall area. For example, when the target partitioning coefficient is 3, the target roadway sidewall area is divided into 3 sidewall sub-areas; when the target partitioning coefficient is 10, the target roadway sidewall area is divided into 10 sidewall sub-areas; when the target partitioning coefficient is 12, the target roadway sidewall area is divided into 12 sidewall sub-areas. The above are just examples and are not limited here.

[0055] The preset division coefficient list is a pre-created list that includes multiple different preset total areas and the preset division coefficients corresponding to each preset total area. Subsequently, the division coefficients can be determined based on the size of the area to be sprayed, dividing the target tunnel sidewall area into a corresponding number of sidewall sub-regions according to the appropriate division coefficients. This effectively controls the progress and timeliness of the shotcreting operation in each sidewall sub-region, avoiding the impact of excessively large sidewall sub-regions on the shotcreting effect. It enables precise shotcreting operations in each sidewall sub-region. Furthermore, it facilitates subsequent shotcreting effect testing of each sidewall sub-region, allowing for rapid location of areas requiring shotcreting repair and improving the efficiency of the entire shotcreting operation process.

[0056] Specifically, the target roadway sidewall region can first be determined based on the initial roadway sidewall image. For example, the initial roadway sidewall image can be input into the trained sidewall region recognition model to output the location information of the target roadway sidewall region. The trained sidewall region recognition model is trained by minimizing the difference between the predicted location information output from the sample initial roadway sidewall image and the sample location information. Other methods can also be used to calculate the corresponding target roadway sidewall region based on the initial roadway sidewall image, which are not limited here.

[0057] Next, determine the total target area corresponding to the target tunnel sidewall region. For example, the total target area can be obtained by converting the region's pixel area. Specifically, the region's pixel area (which can be represented by the number of pixels) can be obtained, along with the camera intrinsic parameters and shooting distance of the spraying robot's camera component. The conversion relationship between physical dimensions and pixels is calculated based on the vacuum camera model and shooting distance. For example, physical width (mm) / pixel = (shooting distance D (mm) × pixel width dx (mm)) / focal length f (mm), physical height (mm) / pixel = (shooting distance D (mm) × pixel height dy (mm)) / focal length f (mm). Typically, pixels are square (dx = dy), so only a conversion factor needs to be calculated. For example, assuming the calculated result is 0.5mm / pixel, this means that one pixel in the image corresponds to a length of 0.5 millimeters in the real world. Furthermore, the region's pixel area, physical width, and physical height are multiplied together to obtain the total target area.

[0058] Furthermore, after obtaining the target total area corresponding to the target roadway sidewall region, the corresponding target division coefficient can be retrieved from the preset division coefficient list based on the target total area. Finally, according to the target division coefficient, the target roadway sidewall region is divided into multiple sidewall sub-regions. Thus, determining the number of sidewall sub-regions based on the target total area corresponding to the target roadway sidewall region effectively controls the progress and timeliness of shotcreting operations in each sidewall sub-region, avoiding the impact of excessively large sidewall sub-regions on the shotcreting effect. This allows for refined shotcreting operations in each sidewall sub-region. Additionally, it facilitates subsequent shotcreting effect testing of each sidewall sub-region, enabling rapid location of areas requiring shotcreting repair and improving the efficiency of the entire shotcreting operation process.

[0059] For example, Figure 5 This is an example diagram illustrating a sidewall sub-region division scenario provided in an embodiment of this application, combined with... Figure 5 As shown, assuming the sidewalls of the surrounding alleyway at the target location point are in the shape of a ring-shaped arch, and the target alleyway sidewall region is a 270-degree arch, this target alleyway sidewall region is divided into 12 parts, resulting in 12 sidewall sub-regions. The above is just an example.

[0060] In some implementations, the corresponding area spraying control parameters can be generated by combining the area wall smoothness and area coordinate information of each sidewall sub-region. For example, step 103, "generating multiple area spraying control parameters based on multiple sidewall sub-regions," may include: identifying the area wall smoothness corresponding to each sidewall sub-region based on the initial tunnel sidewall image; determining the spraying pressure parameter for each sidewall sub-region based on the area wall smoothness; determining the area coordinate information corresponding to each sidewall sub-region; and generating multiple area spraying control parameters by combining the area coordinate information and spraying pressure parameter corresponding to each sidewall sub-region.

[0061] The wall smoothness of this region can refer to the surface smoothness, evenness, or flatness of the corresponding sidewall sub-region. It's understood that the inner wall of the target tunnel may have a rougher construction process during excavation or construction; for example, the surface of the inner wall of the target tunnel may be uneven, resulting in a lower wall smoothness. Each region's wall smoothness only represents the smoothness of its corresponding sidewall sub-region. When calculating the region's wall smoothness, the area of ​​the sub-region and the area of ​​any concave areas within it can be determined. When a sidewall sub-region contains multiple concave areas, the area of ​​each concave area is calculated, and these multiple areas are combined to obtain the total concave area. Then, the region's wall smoothness is obtained based on the ratio between the concave area and the sub-region area. Furthermore, the target sum of the concave and convex areas can be determined by combining the area of ​​the convex area. The region's wall smoothness is then obtained based on the ratio between the target sum and the sub-region area.

[0062] The spraying pressure parameter can be the spraying pressure control value when the spraying assembly performs spraying operations on the corresponding side wall sub-area.

[0063] The region coordinate information refers to the coordinate position of the corresponding sidewall sub-region on the sidewall of the roadway surrounding the current target location point in the target roadway. The region coordinate information may include the upper left and lower right corner coordinates of the corresponding sidewall sub-region, or it may include the four coordinates of upper left, lower left, upper right, and lower right. It may also include the edge line information between the coordinate points. These coordinates can be world coordinates, which are not limited here.

[0064] Specifically, firstly, the smoothness of the wall surface of each sub-region of the tunnel is evaluated based on the initial tunnel sidewall image. The focus is on examining each pre-defined sub-region of the sidewall. For example, the smoothness of the wall surface of each sub-region of the sidewall can be determined by image analysis algorithms (such as calculating the gradient and variance of the image or using a specialized texture analysis model). If the initial tunnel sidewall image looks delicate and uniform with small pixel value variations (such as a newly sprayed concrete surface), the smoothness is high. If the initial tunnel sidewall image looks rough, grainy, or pitted with large pixel value variations (such as an untreated rock surface, cracked or pitted concrete), the smoothness is low. Then, the spraying pressure parameters are determined based on the smoothness of the regional wall surface. Preset rules or models can be defined, establishing the correspondence between "regional wall surface smoothness" and "spraying pressure parameters." For example, the rougher the wall surface, the greater the required spraying pressure to ensure better paint adhesion and filling; conversely, the smoother the wall surface, the lower the required pressure to prevent paint accumulation or dripping. Specifically, the corresponding spraying pressure parameters can be obtained by querying the preset spraying pressure parameter list based on the regional wall surface smoothness. This list contains multiple preset spraying pressure parameters and the corresponding preset regional wall surface smoothness for each parameter. Next, the regional coordinate information of each sidewall sub-region is determined. This can be done by first determining the image coordinates of each sidewall sub-region in the initial tunnel sidewall image, such as determining the pixel coordinates of the upper left and lower right corners of the sub-region in the image. Then, through camera calibration or by combining LiDAR data, the pixel coordinates are converted into real-world 3D coordinates (e.g., a range from X meters from the tunnel start point to Y meters to Z meters in height), obtaining the upper left and lower right world coordinates of each sidewall sub-region. Finally, by combining the area coordinate information and spraying pressure parameters corresponding to each sidewall sub-region, multiple area spraying control parameters are generated. For example, for each sidewall sub-region, the area coordinate information tells the spraying robot where to spray, while the spraying pressure parameters tell the spraying robot how much pressure to use. Therefore, the area coordinate information and spraying pressure parameters can be combined, for example, represented in array form as "[area coordinate information: spraying pressure parameters]", to obtain the area spraying control parameters corresponding to each sidewall sub-region. In addition, these area spraying control parameters can also include spraying pressure control values, spraying rod movement speed, movement position range, etc. By setting independent spraying pressure parameters for different sidewall sub-regions, the spraying assembly can be adaptively controlled to perform spraying operations on different sidewall sub-regions, so as to achieve better spraying results for that sidewall sub-region.

[0065] For example, suppose the target tunnel sidewall area is divided into three sub-regions: A, B, and C. Analyzing the smoothness of these three sub-regions, region A is very rough with a smoothness score of 20, region B is moderate with a smoothness score of 60, and region C is very smooth with a smoothness score of 95. Determine the spraying pressure parameters: 0.9 MPa for region A, 0.5 MPa for region B, and 0.2 MPa for region C. Determine the region coordinates: region A's coordinates are (5.0m-6.0m from the starting point, 1.0m-2.0m height), region B's coordinates are (6.0m-7.0m from the starting point, 1.0m-2.0m height), and region C's coordinates are (7.0m-8.0m from the starting point, 1.0m-2.0m height). Generate zone spraying control parameters: Zone A: {Coordinates: (5.0-6.0m, 1.0-2.0m), Pressure: 0.9MPa}; Zone B: {Coordinates: (6.0-7.0m, 1.0-2.0m), Pressure: 0.5MPa}; Zone C: {Coordinates: (7.0-8.0m, 1.0-2.0m), Pressure: 0.2MPa}. Upon receiving these parameters, the spraying robot will spray Zone A with a pressure of 0.9MPa, then move to Zone B with a pressure of 0.5MPa, and so on, achieving precise and adaptive spraying of the entire tunnel sidewall.

[0066] Using the above method, the target roadway sidewall region can be identified based on the initial roadway sidewall image. The target roadway sidewall region can then be divided into multiple sub-regions for shotcreting operations. Corresponding spraying control parameters are then generated for each sub-region, allowing for subsequent shotcreting operations based on these parameters. This shotcreting planning method, which divides the roadway into multiple sub-regions, not only enables adaptive shotcreting operations for different sub-regions—for example, different shapes and curved surfaces can be targeted with different shotcreting operations, such as varying shotcreting pressure parameters or different shotcreting boom movement speeds—achieving refined shotcreting operations for each sub-region, but also facilitates subsequent shotcreting effect detection for each sub-region, enabling rapid location of areas requiring shotcreting repair and improving the efficiency of the entire shotcreting process.

[0067] 104. Based on the spraying control parameters of multiple areas, control the spraying assembly to spray multiple sidewall sub-areas in sequence, and collect images of the surrounding roadway sidewalls of the target location point after spraying.

[0068] In this embodiment of the application, after obtaining the area spraying control parameters corresponding to each sidewall sub-region, the spraying operation sequence for each sidewall sub-region can be determined, for example, by combining... Figure 5 As shown, assuming there are 12 sidewall sub-regions, the spraying operation starts from the first sidewall sub-region from left to right. Based on multiple area spraying control parameters, the spraying operation is first performed on the first sidewall sub-region according to the area spraying control parameters corresponding to the first sidewall sub-region. For example, the spraying assembly is controlled according to the first area spraying control parameters so that the spraying assembly performs the spraying operation on the first sidewall sub-region. When the spraying operation of the first sidewall sub-region is completed, the next step is taken, and the spraying operation of the second sidewall sub-region is performed according to the area spraying control parameters corresponding to the second sidewall sub-region. This process continues until the spraying operation of the 12th sidewall sub-region is completed. Subsequently, in order to ensure the shotcreting effect of each sidewall sub-area, and to check for omissions and make up for deficiencies, an image recognition method can be used to collect images of the target roadway sidewall after shotcreting, around the current target location point. This allows for the identification of the effect of the first round of shotcreting based on the target roadway sidewall image, so that defective areas can be repaired separately and shotcreting can be carried out to improve the shotcreting effect.

[0069] In some implementations, before step 104, the method further includes: inputting the initial tunnel sidewall image into a pre-trained target image recognition model to obtain an initial recognition result, and determining the range of the work area to be sprayed based on the initial recognition result. Then step 104 may include: combining the spraying area range and multiple area spraying control parameters, controlling the spraying assembly to sequentially spray multiple sidewall sub-regions, and acquiring images of the surrounding tunnel sidewalls of the target location point after spraying.

[0070] It should be noted that, in order to improve the efficiency of the spraying operation and minimize unnecessary spraying operations, the area to be sprayed can be identified based on the initial tunnel sidewall image. This area refers to the actual range of the sidewall area requiring spraying at the current target location point in the target tunnel. This area may include multiple area coordinates or other location information representing the area to be sprayed; no limitation is made here. Furthermore, by combining the actual area of ​​the sidewall to be sprayed with multiple area spraying control parameters, the spraying assembly is controlled to sequentially spray multiple sidewall sub-areas. It should be noted that when a sub-area of ​​the tunnel sidewall does not require spraying, the corresponding area spraying control parameter can be "0" or other command parameters indicating refusal to execute the spraying operation, such as "skip spraying" or "complete spraying" command parameters. This achieves precise spraying of the sidewall area, saving unnecessary spraying operations and improving the efficiency of the spraying operation.

[0071] It should be noted that this pre-trained target image recognition model can be trained by minimizing the difference between the predicted recognition results output from the sample target tunnel sidewall image and the sample recognition results. This training process will be described later.

[0072] By using the above method, the sequence of shotcreting operations for each sidewall sub-region can be determined. Based on the area spraying control parameters corresponding to each sidewall sub-region, the shotcreting components are controlled to sequentially perform shotcreting operations on multiple sidewall sub-regions. After all sidewall sub-regions have been shotcreted, images of the target roadway sidewall surrounding the target location point after shotcreting are acquired. This allows for subsequent identification of the first round of shotcreting operation effects based on the target roadway sidewall images, enabling subsequent individual shotcreting repair of defective areas and improving the shotcreting operation effect.

[0073] 105. Based on the target tunnel sidewall image, determine the shotcrete completion status of multiple sidewall sub-regions.

[0074] In this embodiment, after obtaining an image of the target tunnel sidewall after the grouting status of the current target location, the grouting completion status of multiple sidewall sub-regions corresponding to the current target location can be identified based on this image. This grouting completion status refers to the degree of completion of grouting on the surrounding tunnel sidewalls (i.e., multiple sidewall sub-regions) of the target location. For example, the grouting coverage rate and coating thickness of the wall can be determined from the target tunnel sidewall image, thereby determining the grouting completion status of the multiple sidewall sub-regions. This allows for subsequent decisions based on the grouting completion status to determine whether further grouting repairs are needed on the surrounding tunnel sidewalls of the target location, such as deciding whether grouting is ultimately required or not, thereby improving the grouting operation effect on the target tunnel.

[0075] The shotcreting completion status can be either shotcreting completed or shotcreting incomplete, that is, either of the two states. The specific state can be determined according to the actual situation. For example, it can be determined by identifying the shotcreting effect of multiple sidewall sub-regions based on the target tunnel sidewall image. For example, when the shotcreting effect of all sidewall sub-regions reaches the expected level, the shotcreting completion status is determined to be shotcreting completed. Conversely, when there is at least one sidewall sub-region whose shotcreting completion status is shotcreting incomplete, the shotcreting completion status is shotcreting incomplete.

[0076] In this application, a target image recognition model can be pre-trained to identify the shotcrete status of the tunnel sidewall. For example, the training process of the pre-trained target image recognition model is as follows: First sample images of the tunnel sidewall without shotcrete and second sample images of the tunnel sidewall with shotcrete are acquired. The first sample image is labeled as a region requiring shotcrete, and the second sample image is labeled as a region not requiring shotcrete. The first sample image is input into the pre-trained image recognition model to obtain a first predicted recognition result. A first loss is constructed based on the difference between the first predicted recognition result and the first sample recognition result. The network parameters of the pre-trained image recognition model are adjusted based on the first loss to initialize the image recognition model. The second sample image is input into the initial image recognition model to obtain a second predicted recognition result. A second loss is constructed based on the difference between the second predicted recognition result and the second sample recognition result. The network parameters of the initial image recognition model are adjusted based on the second loss to obtain the pre-trained target image recognition model.

[0077] For example, wall visual images can be collected under conditions where lighting and environmental factors are consistent with the working environment. These wall visual images are divided into two categories: one category is wall images where spraying has not been completed, i.e., images of the first tunnel sidewall (for example, this first tunnel sidewall image may be an image with exposed or partially exposed metal steel mesh, depending on the actual scenario, and is not limited here); the other category is wall images where spraying has been completed, i.e., images of the second tunnel sidewall (for example, this second tunnel sidewall image may be a wall with cement slurry attached, with no exposed steel mesh). At least 1000 wall visual images are collected under various conditions according to these two categories, and each wall visual image is labeled, with corresponding label data for training a target image recognition model.

[0078] In some implementations, the shotcrete status of the corresponding tunnel sidewall can be identified using a trained target image recognition model. For example, step 105 may include: applying a pre-trained target image recognition model to the target tunnel sidewall image to obtain a recognition result; wherein the pre-trained target image recognition model is trained by minimizing the difference between the predicted recognition result output by the sample target tunnel sidewall image and the sample recognition result; and determining the shotcrete completion status corresponding to the plurality of sidewall sub-regions based on the recognition result.

[0079] Specifically, the target tunnel sidewall image can be input into a pre-trained target image recognition model, enabling the model to perform recognition based on the image. Since this model is pre-trained using an end-to-end model detection algorithm, with pre-sampled feature datasets manually labeled and then trained, it can be used for real-time spraying area recognition. Therefore, the model can be directly used to identify the spraying status based on the input target tunnel sidewall image. Finally, the spraying completion status of multiple sidewall sub-regions is determined based on the recognition results. For example, the spraying completion status of each sub-region can be determined individually based on the recognition results. For instance, assuming six sub-regions are divided, the first, second, third, fourth, and fifth sub-regions are sprayed and do not require further spraying, while the sixth sub-region is not yet sprayed and requires spraying. This is just an example. In this way, the shotcreting completion status of multiple sidewall sub-areas can be determined, and a decision can be made on whether shotcreting repair is needed for multiple sidewall sub-areas based on the shotcreting completion status, thereby improving the shotcreting operation effect for the target roadway.

[0080] In this embodiment, the similarity can be determined by comparing the initial roadway sidewall image before shotcreting with the target roadway sidewall image after shotcreting. Based on this similarity, the shotcreting completion status of multiple sidewall sub-regions at the target location can be determined. This allows for subsequent decisions based on the shotcreting completion status to determine whether further shotcreting repairs are needed on the roadway sidewalls surrounding the target location, thereby improving the shotcreting operation effect for the target roadway.

[0081] In some implementations, a pre-trained target image recognition model can output recognition results based on the target tunnel sidewall image and the initial tunnel sidewall image to determine the shotcrete completion status of multiple sidewall sub-regions according to the recognition results. For example, step 105 may include: (105.A.1) Input the target tunnel sidewall image and the initial tunnel sidewall image into the pre-trained target image recognition model for recognition, and obtain the recognition result; (105.A.2) Based on the recognition results, determine the contrast similarity between the target tunnel sidewall image and the initial tunnel sidewall image; (105.A.3) When the comparison similarity is less than or equal to the preset similarity threshold, the shotcreting completion status of multiple sidewall sub-regions is determined as shotcreting completion; (105.A.4) When the comparison similarity is greater than the preset similarity threshold, the shotcreting completion status of multiple side wall sub-regions is determined as shotcreting incomplete.

[0082] The pre-trained target image recognition model is trained by minimizing the difference between the predicted recognition results output from the sample target tunnel sidewall image and the sample initial tunnel sidewall image. For example, a large number of "before spraying photos" and "after spraying photos" and their corresponding "standard answers" (i.e., sample recognition results, such as sample comparison similarity) can be used as training data. Based on this training data, the model continuously tries to compare the sample images and compares the predicted results with the "standard answers," continuously adjusting the model's internal parameters until its predicted results are very close to the standard answers. In this way, the target image recognition model is trained.

[0083] The identification result may include comparison similarity and / or sub-similarity of each sidewall sub-region before and after spraying. Through the identification result, the state of multiple sidewall sub-regions before and after spraying can be determined.

[0084] The comparison similarity can be the similarity between the target tunnel sidewall image and the initial tunnel sidewall image. Specifically, it can be the pixel similarity between the target tunnel sidewall image and the initial tunnel sidewall image. The comparison can be made between the pixels of the same content location on the target tunnel sidewall area and the initial tunnel sidewall image, and the number of pixels at the same location can be counted. The total number of pixels on the target tunnel sidewall area can also be determined. The comparison similarity is determined based on the ratio between the number of pixels at the same location and the total number of pixels.

[0085] The preset similarity threshold can be a value used to determine the pixel similarity between the target tunnel sidewall image and the initial tunnel sidewall image. It is understood that since the target tunnel sidewall area (multiple sidewall sub-areas) is generally inconsistent before and after shotcreting, for example, the wall color after shotcreting is different from the wall color before shotcreting, the pixel similarity before and after shotcreting is usually low, close to 0. Therefore, the preset similarity threshold can be set to 0, or the preset similarity threshold can be set to below 0.1, depending on the actual situation, and is not limited here.

[0086] Specifically, the target tunnel sidewall image and the initial tunnel sidewall image can be input together into a pre-trained target image recognition model. The pre-trained model then performs recognition based on these two images to obtain the recognition result. Next, based on the recognition result, the contrast similarity between the target tunnel sidewall image and the initial tunnel sidewall image is determined. Further, this contrast similarity is compared with a preset similarity threshold. If the contrast similarity is less than or equal to the preset threshold (e.g., assuming the threshold is 0.05 and the similarity is 0.01), it indicates that the pixel similarity between the target tunnel sidewall image and the initial tunnel sidewall image is relatively low, meaning the two images are not similar, which matches the expected effect before and after shotcreting. Therefore, the shotcreting completion status corresponding to multiple sidewall sub-regions at the target location point is determined as shotcreting completion. On the other hand, if the comparison similarity is greater than a preset similarity threshold—for example, a comparison similarity of 0.1, which is greater than the preset similarity threshold of 0.05—it indicates that some wall surfaces in multiple sidewall sub-regions are not completely coated or are incomplete. This results in some pixels being identical between the two images, leading to a high comparison similarity. Therefore, the spraying completion status of multiple sidewall sub-regions corresponding to the target location point is determined as incomplete spraying. In this way, the spraying completion status of multiple sidewall sub-regions can be determined, allowing for a decision on whether to perform spraying repairs on these sub-regions, thereby improving the effectiveness of spraying operations on the target tunnel.

[0087] In some implementations, when the shotcreting completion status corresponding to multiple sidewall sub-regions is determined to be shotcreting incomplete, a corresponding target similar region can be identified from the multiple sidewall sub-regions after shotcreting, and shotcreting repair can be performed on the target similar region. An image of the surrounding roadway sidewall at the target location point after shotcreting repair can be acquired to further evaluate the shotcreting completion status based on the image of the sidewall to be identified. For example, after step 105, step "determine the shotcreting completion status corresponding to the multiple sidewall sub-regions based on the identification result" or step (105.A.4), if the shotcreting completion status is identified as shotcreting incomplete, the following may also be included: (B.1) Identify at least one target similarity region between the target tunnel sidewall image and the initial tunnel sidewall image; (B.2) Generate corresponding target area spraying control parameters based on the target coordinate information of similar target areas; (B.3) Control the spraying assembly to spray and repair the corresponding similar areas on the sidewalls around the target location point according to the spraying control parameters of the target area; (B.4) Collect images of the sidewalls of the surrounding roadway at the target location after shotcrete repair; (B.5) Based on the images of the sidewalls to be identified, re-evaluate the shotcrete completion status of multiple sidewall sub-regions.

[0088] The target similarity region can be a region with the same or similar pixels between the target tunnel sidewall image and the initial tunnel sidewall image. The target similarity region can be one of the multiple sidewall sub-regions, or it can be a local region within the sidewall sub-regions. No limitation is made here.

[0089] The target coordinate information can be the location information of a similar region to the target. It can be the coordinates of the upper left corner and the lower right corner of the similar region. These can also be world coordinates. For details, please refer to the description of "region coordinate information" above. It will not be repeated here.

[0090] The target area spraying control parameters refer to the spraying control parameters for similar areas of the target. For an explanation of the target area spraying control parameters, please refer to the previous "Area Spraying Control Parameters". They will not be repeated here.

[0091] The sidewall image to be identified can be a sidewall image after shotcrete repair of the sidewall of the roadway surrounding the target location has been completed, and it contains information about the range of the sidewall area of ​​the roadway surrounding the target location.

[0092] It should be noted that after determining that the shotcreting completion status of multiple sidewall sub-regions is incomplete, shotcreting repair is required for the missing areas. Specifically, in the target tunnel sidewall image after shotcreting, some areas are highly similar to the initial image before shotcreting (e.g., before shotcreting, there was a rough rock surface or metal mesh (the solidified mesh before shotcreting), and after shotcreting, these areas are still rough rock surfaces or exposed metal mesh, with little difference before and after shotcreting, indicating that the shotcreting was not applied or was too thin). Therefore, based on the "identification results" (e.g., difference heatmap, pixel-level comparison report), these "areas similar to the initial image" are automatically circled, and the specific range of each area is determined. For example, the comparison similarity can include the sub-similarity of each sidewall sub-region before and after shotcreting. Based on the sub-similarity of each sidewall sub-region, at least one target similar region between the target tunnel sidewall image and the initial tunnel sidewall image is determined. For example, sidewall sub-regions with a sub-similarity greater than a preset sub-similarity threshold are selected as target similar regions. Thus, the areas where the target location point still needs shotcreting repair after the first round of shotcreting are identified.

[0093] Then, based on the target coordinate information of the similar target areas, the corresponding target area spraying control parameters are generated. For example, after determining the similar target areas, the target coordinate information of these areas can be determined. The method for determining this "target coordinate information" can refer to the process for determining "area coordinate information," and is not limited here. Furthermore, after obtaining the target coordinate information of the similar target areas, the corresponding target area spraying control parameters can be generated based on the target coordinate information. In addition, the target area spraying control parameters can also be generated by combining the spraying pressure parameters. For details, refer to the previous process for determining "area spraying control parameters," which will not be elaborated here.

[0094] Furthermore, the spraying assembly is controlled according to the target area spraying control parameters to spray and repair the corresponding similar areas on the sidewalls surrounding the target location. The spraying assembly first moves to the starting position of the similar area and then sprays evenly along the area according to the set pressure and flow rate (e.g., layer by layer from left to right and from top to bottom), ensuring that the repaired area reaches the qualified thickness (e.g., 30mm) and has a smooth surface. The entire process requires no manual intervention, and the system monitors the position and process parameters of the spraying assembly in real time to avoid missed spraying or repeated spraying. After the spraying and repair of the similar area is completed, the spraying robot controls the camera assembly to acquire images of the sidewalls of the surrounding tunnel around the target location after the spraying and repair. Finally, based on the images of the sidewalls to be identified, the spraying completion status of multiple sidewall sub-regions is re-evaluated. In this way, by detecting the areas that need repair and performing spraying operations on them, the spraying operation effect is improved, and the reliability is enhanced.

[0095] In some implementations, the image of the sidewall to be identified can be compared with the initial roadway sidewall image to determine a first similarity. If the first similarity is greater than a preset similarity threshold, the image of the sidewall to be identified and the target roadway sidewall image can be compared to determine a second similarity. If the second similarity is less than or equal to a preset spraying similarity threshold, the spraying completion status corresponding to multiple sidewall sub-regions is determined to be spraying completed; otherwise, the spraying completion status corresponding to multiple sidewall sub-regions is determined to be spraying incomplete. For example, step (B.5) may include: inputting the sidewall image to be identified into a pre-trained target image recognition model for recognition to obtain a first recognition result; when it is determined that spraying is required based on the first recognition result, determining a first similarity between the sidewall image to be identified and the initial tunnel sidewall image; when the first similarity is greater than a preset similarity threshold, determining a second similarity between the sidewall image to be identified and the target tunnel sidewall image; when the second similarity is less than or equal to a preset spraying similarity threshold, determining the spraying completion status of multiple sidewall sub-regions as spraying completed; when the second similarity is greater than the preset spraying similarity threshold, determining the spraying completion status of multiple sidewall sub-regions as spraying incomplete.

[0096] The explanations of the first and second identification results can be found in the previous description of "Identification Results", and the explanations of the first and second similarities can be found in the previous description of "Comparison Similarity". They will not be repeated here.

[0097] Specifically, after obtaining the image of the sidewall to be identified, it can be input into a pre-trained target image recognition model. This model then performs recognition based on the sidewall image, outputting a first recognition result. When the first recognition result determines that multiple sidewall sub-regions require spraying, to improve recognition accuracy, the pixel similarity between the sprayed sidewall image and the initial tunnel sidewall image before spraying can be further determined, i.e., the first similarity. This first similarity is then compared to a preset similarity threshold. If the first similarity is less than or equal to the preset similarity threshold, it indicates that the pixels of the surrounding sidewall region at the target location point are significantly different from those before spraying, and the spraying repair of the similar target region has achieved the expected effect. Therefore, the spraying completion status corresponding to multiple sidewall sub-regions is determined as spraying completion.

[0098] On the other hand, if the first similarity is greater than the preset similarity threshold, it means that some wall surfaces in multiple sidewall sub-regions have not been sprayed completely or are missing, resulting in some pixels being the same between the two images and the comparison similarity being too high. However, since it has already undergone a spraying repair, there may be other factors that cause the image pixels to not achieve the expected effect. In order to avoid repeatedly repairing the same area and improve the efficiency of the spraying operation of the entire tunnel, the pixel comparison can be performed between the target tunnel sidewall image after the first spraying and the current sidewall image to be identified. For example, the sidewall image to be identified and the target tunnel sidewall image can be input together into the pre-trained target image recognition model so that the pre-trained target image recognition model can perform recognition based on the sidewall image to be identified and the target tunnel sidewall image to output a second recognition result. The pixel similarity between the sidewall image to be identified after spraying repair and the target tunnel sidewall image after the first spraying is determined based on the second recognition result, i.e., the second similarity.

[0099] Furthermore, the second similarity is compared with a preset spraying similarity threshold. This preset spraying similarity threshold can be a judgment value for the pixel similarity between images after spraying repair. Since the surrounding tunnel sidewall area after spraying the target location point generally shows a large amount of slurry color, even after repair, they are still relatively similar. Therefore, the similarity between the sidewall image to be identified and the target tunnel sidewall image after the first spraying is relatively high. Therefore, the preset spraying similarity threshold can be set to 0.9, 0.95, 0.99, etc., or it can be determined based on a preset similarity threshold. For example, the sum of the preset similarity threshold and the preset spraying similarity threshold is 1. When the preset similarity threshold is 0.05, the preset spraying similarity threshold is 0.95. Therefore, on the one hand, if the second similarity is less than or equal to the preset spraying similarity threshold, it means that the target similar area has been sprayed repaired, so there is a pixel difference. Therefore, the spraying completion status corresponding to multiple sidewall sub-regions is determined as spraying completion. It should be noted that, since the repair area is relatively small, the pixel difference is relatively small, which is consistent with the logic of consumer repair. It also indicates that the difference may be caused by other factors (such as illumination or other factors). Therefore, if it is not particularly serious, there is no need to perform spraying repair again to avoid recognition errors due to lighting issues, and to improve efficiency.

[0100] On the other hand, if the second similarity score is greater than the preset spraying similarity threshold, it indicates that there are still significant pixel differences after spraying repair of the target similar areas. In this case, the spraying completion status of multiple sidewall sub-regions is determined to be incomplete, and subsequent spraying repair processes need to be performed on multiple sidewall sub-regions. The spraying repair process is described above and will not be repeated here. Therefore, by following the above method, the recognition results of the spraying status can be further enhanced to avoid errors caused by unilateral detection, improve the accuracy of the recognition results, and thus improve the effect of subsequent spraying operations.

[0101] In some implementations, step (B.5) may include: dividing the sidewall image to be identified into multiple sub-regions to obtain images containing multiple sub-regions to be identified; inputting each sub-region image to be identified into a pre-trained target image recognition model to obtain a sub-recognition result; and determining the spraying completion sub-state corresponding to each sidewall sub-region based on each sub-recognition result, where each spraying completion sub-state can be either "the sub-region needs spraying" or "the sub-region does not need spraying". The pre-trained target image recognition model can be trained by minimizing the difference between the predicted recognition result output from the sample target tunnel sidewall image and the sample recognition result.

[0102] For example, an image of the sidewall to be identified is acquired within a 270-degree area of ​​the sidewall of the tunnel at the target location. This image is then divided into sub-regions, specifically 12 sub-regions, each with an angle range of 22.5 degrees. A target image recognition model is then used to identify the corresponding sub-region image. By using each sub-region as a unit for spraying and repair work, and by individually identifying and detecting the spraying effect of each sub-region, the location of poorly sprayed areas on the tunnel sidewall can be quickly identified, and targeted spraying and repair work can be carried out, simplifying the spraying process and improving spraying efficiency.

[0103] In this embodiment, image recognition and lidar can also be combined to determine the shotcrete completion status of multiple sidewall sub-regions, so as to improve the accuracy of the shotcrete completion status.

[0104] In some implementations, a pre-trained target image recognition model outputs recognition results based on the target tunnel sidewall image and the initial tunnel sidewall image. It also obtains regional point cloud information corresponding to each sidewall sub-region using LiDAR to determine the smoothness of the regional wall surface. Based on the contrast similarity and regional wall smoothness corresponding to the recognition results, the completion status of the spraying for multiple sidewall sub-regions is determined. For example, step 105 may include: inputting the target tunnel sidewall image into the pre-trained target image recognition model for recognition to obtain the target recognition result; obtaining regional point cloud information corresponding to each sidewall sub-region using depth information from a depth camera or LiDAR, and determining the regional wall smoothness for each sidewall sub-region based on the regional point cloud information; when it is determined according to the target recognition result that spraying is not required, and the regional wall smoothness is greater than or equal to a preset smoothness threshold, the completion status of the spraying for multiple sidewall sub-regions is determined as spraying completed; when it is determined according to the target recognition result that spraying is required, or the regional wall smoothness is less than a preset smoothness threshold, the completion status of the spraying for multiple sidewall sub-regions is determined as spraying incomplete.

[0105] The description of the target recognition result can be found in the previous section on "Recognition Results," and will not be repeated here. Similarly, the description of the target comparison similarity can be found in the previous section on "Comparison Similarity," and will not be repeated here.

[0106] The depth information can be the distance information between objects in the scene and the camera obtained by the depth camera through different principles. The point cloud information of the region can be determined through this depth information.

[0107] The point cloud information in this area can be a dataset composed of discrete three-dimensional coordinate points (x, y, z) obtained by scanning multiple side wall sub-regions of the current target location. These points are densely distributed and accurately restore the geometric features such as the three-dimensional outline, surface undulation, and spatial position of objects (such as the side wall of a lane) in the area. It is the core data carrier for digitally describing the physical morphology of the area.

[0108] The smoothness of the wall surface in this area can refer to the surface smoothness, smoothness, or flatness of the corresponding side wall sub-area. For details, please refer to the previous description, which will not be repeated here.

[0109] Specifically, the target tunnel sidewall image is input into a pre-trained target image recognition model, which then performs recognition based on the target tunnel sidewall image and the initial tunnel sidewall image to output the target recognition result. Furthermore, the target recognition result can be used to preliminarily determine whether spraying is required, including two selectable states: spraying required or not spraying required.

[0110] Then, the point cloud information corresponding to each sidewall sub-region is acquired using LiDAR. For example, the shotcrete robot also includes a LiDAR component. The LiDAR component is controlled to scan the surrounding alleyway sidewall area after shotcreting at the target location point. The LiDAR emits a laser beam and generates "point cloud data" for each sidewall sub-region by receiving reflected signals. This area point cloud information can be understood as reconstructing the three-dimensional outline of each sidewall sub-region after shotcreting using millions of "three-dimensional coordinate points". Further, based on the area point cloud information of each sidewall sub-region, the corresponding wall smoothness is calculated. For example, by calculating indicators such as the height difference and curvature change of adjacent points in the point cloud, if the overall undulation of the point cloud is small and the height difference is uniform, it indicates that the wall surface is flat (high smoothness); if the point cloud has obvious bulges or depressions, it indicates that the wall surface is rough (low smoothness). A quantitative score of "area wall smoothness" (e.g., 0-100 points, the higher the score, the smoother the surface) is output for each sidewall sub-region. In addition, the specific determination process of "area wall smoothness" can also be referred to earlier.

[0111] Furthermore, on the one hand, if the target recognition results determine that spraying is unnecessary, and the smoothness of the area's wall surface is greater than or equal to a preset smoothness threshold, it indicates that the spraying coverage is sufficient, there are no obvious missed areas, and the wall surface after spraying is sufficiently smooth, meeting the quality standards. Therefore, the spraying completion status corresponding to multiple side wall sub-areas is determined as spraying complete. On the other hand, if the target recognition results determine that spraying is necessary, or the smoothness of the area's wall surface is less than a preset smoothness threshold, it indicates that the spraying coverage is insufficient, there are missed areas, or the wall surface after spraying is still rough (such as pits or bumps), failing to meet the quality standards. Therefore, the spraying completion status corresponding to multiple side wall sub-areas is determined as spraying incomplete. In this way, through the dual verification of "image comparison" and "LiDAR smoothness measurement," the shortcomings of a single standard are avoided. For example, looking only at the image may overlook the problem of roughness on the wall surface after spraying, and looking only at the smoothness may overlook the problem of missed spraying. Ultimately, a comprehensive and accurate judgment of the "coverage and smoothness" of the spraying effect is achieved, improving the effect of the spraying operation.

[0112] It should be noted that when multiple sidewall sub-areas are detected as having an "incomplete spraying" status, the remaining incomplete sidewall sub-areas will be sprayed again. After the second spraying, the "spraying completion status identification" will be performed again to confirm whether there are any incomplete sidewall sub-areas. If the detection result still shows incomplete sidewall sub-areas, the spraying operation will be performed again, and so on. It should also be noted that to avoid repeated spraying operations at the same target location, which would affect the spraying progress, a maximum of 3 spraying operations can be set for the same sidewall sub-area at the same target location. If there are sidewall sub-areas that are still incomplete after 3 spraying operations, these are identified as abnormal sidewall sub-areas, and the location information of the abnormal area is recorded, along with an alarm message. Then, the operation process continues to the next target location, i.e., step 106.

[0113] Using the above methods, the shotcrete completion status of multiple sidewall sub-regions corresponding to the current target location can be identified based on the target tunnel sidewall image. This shotcrete completion status refers to the degree of completion of shotcrete application on the surrounding tunnel sidewalls (i.e., multiple sidewall sub-regions). For example, the shotcrete coverage and coating thickness can be determined from the target tunnel sidewall image, thereby determining the shotcrete completion status of multiple sidewall sub-regions. This allows for subsequent decisions based on the shotcrete completion status to determine whether further shotcrete repairs are needed on the surrounding tunnel sidewalls, thus improving the shotcrete operation effect for the target tunnel.

[0114] 106. When the shotcreting completion status is "shotcreting complete", return to the execution to determine the remaining path length corresponding to the target roadway, and stop the shotcreting operation on the target roadway when the corresponding remaining path length is less than or equal to the preset length threshold.

[0115] In this embodiment of the application, after obtaining the spraying completion status, if the spraying completion status is "spraying completed", it means that the spraying operation on the sidewall area of ​​the surrounding roadway at the current target location point in the target roadway is completed. At this time, in order to complete the spraying operation on all sidewall areas of the target roadway, the remaining path length of the target roadway can be further determined, that is, return to step 101 and start from step 101 to execute steps 101 to 104. Specifically, when the remaining path length is greater than a preset length threshold, the motorized component can be controlled to travel to the next target location point in the target roadway and the initial roadway sidewall image of the surrounding roadway sidewall of the target location point can be collected. Based on the initial roadway sidewall image, multiple sidewall sub-regions corresponding to the target location point are determined, and multiple area spraying control parameters are generated based on the multiple sidewall sub-regions. Based on the multiple area spraying control parameters, the spraying component is controlled to spray the multiple sidewall sub-regions in sequence, and the target roadway sidewall image of the surrounding roadway sidewall of the target location point after spraying is collected. Based on the target roadway sidewall image, the spraying completion status corresponding to the multiple sidewall sub-regions is determined. Following the above method, further shotcreting operations are performed on the sidewalls of the surrounding roadways at other locations in the target roadway to ensure that the shotcreting operation can be completed on the sidewalls of the entire target roadway. This achieves automated shotcreting operation on the sidewalls of the entire target roadway, improving the efficiency and effectiveness of shotcreting operations on the sidewalls of the target roadway.

[0116] Using the above method, the shotcreting status of the surrounding roadway sidewalls of each target location can be determined based on the shotcreting completion status of the target location. The shotcreting operation of the surrounding roadway sidewalls of other target locations in the target roadway can be repeated according to the aforementioned process, which improves the degree of automation of shotcreting operation and improves the efficiency and effect of shotcreting operation on the sidewalls of the target roadway.

[0117] To facilitate understanding of the embodiments of this application, the following is a general example of the operation process of the shotcrete robot, wherein... Figure 6 This is an example control flow diagram of the shotcrete robot provided in the embodiments of this application, which can be combined with... Figure 6 The specific process is as follows: The shotcrete robot begins its work; it moves to the entrance of the tunnel where it needs to work; real-time depth RGB information is acquired via a camera, which can be understood as an image of the target tunnel entrance; the host computer (such as a navigation and detection controller, i.e., a processor) determines the robot's working direction (navigation direction, forward direction) and working time based on this image; visual navigation (image recognition) determines the direction of the spraying operation. After confirming the personnel ratio in the above process, the shotcrete operation officially begins. Taking the first step (1-step) of the spraying operation as an example, the specific spraying process is as follows: S1, move 0.3 meters in the direction of the tunnel operation (this is an example, the single forward step length is determined by the size of the working nozzle and the flow rate), which is the "next target location point" at this time; S2. Quickly detect the spraying operation results of areas 1 to 12 within a 270-degree range, and divide them into two results: spraying or not spraying. This can be determined based on the initial tunnel sidewall image of the current target location. S3. If one or more areas need to be sprayed, and the number of sprays is less than 3 (this can be set according to the actual situation, or only 2 sprays are required), proceed to process S4; otherwise, proceed to process S6. S4. Turn on the spraying switch to begin spraying. S5. Complete the spraying operation for areas 1 to 12; S6. Turn off the spraying switch, continue forward 0.3m, and proceed to the next step (i.e., the next target location) of the spraying operation. S7. Determine if the operation time has expired or the navigation has reached the end of the roadway where progress is impossible. If not, return to step S1; if so, navigate back along the historical operation route, report the detection results and records, and issue an alarm to reclaim the operation.

[0118] The above process enables automated shotcrete operation by shotcrete robots, thereby improving shotcrete efficiency.

[0119] As can be seen from the above embodiments, when the shotcrete robot of this application performs shotcrete operation on the target tunnel, it first determines the remaining path length corresponding to the target tunnel. When the remaining path length is greater than a preset length threshold, it controls the motorized component of the shotcrete robot to travel to the next target location point in the target tunnel to acquire an initial tunnel sidewall image of the surrounding tunnel sidewalls of the target location point. Then, based on the initial tunnel sidewall image, it determines multiple sidewall sub-regions corresponding to the target location point, and generates multiple area spraying control parameters for these multiple sidewall sub-regions. Finally, it controls the shotcrete assembly of the shotcrete robot according to these multiple area spraying control parameters. The system performs shotcreting operations on multiple sidewall sub-regions to acquire images of the target roadway sidewalls surrounding the target location after shotcreting. Finally, based on the acquired target roadway sidewall images, it determines the shotcreting completion status corresponding to these multiple sidewall sub-regions. If the shotcreting completion status is "shotcreting complete," it returns to the step of determining the remaining path length corresponding to the target roadway and continues to perform shotcreting operations on the sidewalls surrounding the next target location in the target roadway until the remaining path length is less than or equal to a preset length threshold. This is considered as the shotcreting of all roadway sidewalls of the target roadway being completed. In this way, automatic shotcreting operations on the target roadway are achieved.

[0120] Therefore, compared to the inefficient manual grouting method in related technologies, which is not only inefficient but also poses a safety hazard to workers under high dust conditions, and the inefficient method of using remote-controlled robots to replace manual labor, which requires human operation and thus lacks flexibility and affects the progress of grouting operations, this application uses a spraying robot to determine the remaining path length of the target tunnel, thereby determining the progress of the spraying operation. Through image recognition, based on the initial tunnel sidewall images collected at each target location, multiple sidewall sub-regions requiring spraying are identified, generating multiple area spraying control parameters for spraying control. The spraying completion status is identified based on the collected target tunnel sidewall images. When the spraying of the tunnel sidewall area at a target location is completed, the robot continues to the next target location to perform the spraying operation until completion. Thus, there is no need for manual grouting or human operation of the remote-controlled robot; instead, the spraying robot performs adaptive spraying operations on the target tunnel, offering flexibility and convenience, improving the efficiency of the spraying operation, reducing operation time, and increasing the progress of the spraying operation.

[0121] To facilitate better implementation of the shotcrete robot control method provided in this application, this application also provides a control device based on the shotcrete robot described above. The meanings of the terms used are the same as in the shotcrete robot control method described above, and specific implementation details can be found in the descriptions in the method embodiments.

[0122] Please seeFigure 7 , Figure 7 This is a schematic diagram of the control device for a shotcrete robot provided in an embodiment of this application. The control device for the shotcrete robot is integrated into the computer equipment of this application. If applied to a shotcrete robot, the shotcrete robot includes a shotcrete assembly and a motor assembly. The control device for the shotcrete robot may include a determining unit 401, a first control unit 402, a generating unit 403, a second control unit 404, a status detection unit 405, and a repeating execution unit 406.

[0123] Determining unit 401 is used to determine the remaining path length corresponding to the target roadway; The first control unit 402 is used to control the motor component to travel to the next target location point in the target tunnel when the remaining path length is greater than a preset length threshold, and to collect the initial tunnel sidewall image of the surrounding tunnel sidewall of the target location point. The generation unit 403 is used to determine multiple sidewall sub-regions corresponding to the target location points based on the initial roadway sidewall image, and to generate multiple area spraying control parameters based on the multiple sidewall sub-regions. The second control unit 404 controls the spraying assembly to spray multiple sidewall sub-regions sequentially according to multiple area spraying control parameters, and collects images of the surrounding roadway sidewalls of the target location point after spraying. The status detection unit 405 is used to determine the shotcreting completion status of multiple sidewall sub-regions based on the target roadway sidewall image; The repeat execution unit 406 is used to return to determine the remaining path length corresponding to the target roadway when the shotcreting completion status is shotcreting complete, and stop the shotcreting operation on the target roadway when the corresponding remaining path length is less than or equal to a preset length threshold.

[0124] In some embodiments, the generating unit 403 is further configured to: The target roadway sidewall region is determined based on the initial roadway sidewall image, and the total target area corresponding to the target roadway sidewall region is determined. From the list of preset division coefficients, query the target division coefficient corresponding to the target total area. The list of preset division coefficients includes multiple different preset total areas and the preset division coefficient corresponding to each preset total area. The target roadway sidewall region is divided according to the target division coefficient, resulting in multiple sidewall sub-regions.

[0125] In some embodiments, the state detection unit 405 is further configured to: The target tunnel sidewall image and the initial tunnel sidewall image are input into the pre-trained target image recognition model for recognition, and the recognition result is obtained. Among them, the pre-trained target image recognition model is obtained by minimizing the difference between the predicted recognition results output by the sample target tunnel sidewall image and the sample initial tunnel sidewall image. Based on the recognition results, the contrast similarity between the target tunnel sidewall image and the initial tunnel sidewall image is determined; When the comparison similarity is less than or equal to the preset similarity threshold, the shotcreting completion status of multiple sidewall sub-regions is determined as shotcreting completion. When the comparison similarity is greater than the preset similarity threshold, the shotcreting completion status of multiple sidewall sub-regions is determined as shotcreting incomplete.

[0126] In some embodiments, the control device for the shotcrete robot further includes a third control unit for: Based on the recognition results, at least one target similarity region is determined between the target tunnel sidewall image and the initial tunnel sidewall image; Based on the target coordinate information of similar target areas, generate corresponding target area spraying control parameters; According to the target area spraying control parameters, the spraying assembly is controlled to spray and repair the corresponding target-similar area on the side wall around the target location point; Collect images of the sidewalls of the surrounding tunnels at the target location after shotcrete repair; Based on the images of the sidewalls to be identified, the shotcrete completion status of multiple sidewall sub-regions is reassessed.

[0127] In some implementations, the third control unit is further configured to: The image of the sidewall to be identified and the initial image of the roadway sidewall are input into the pre-trained target image recognition model for recognition, and the first recognition result is obtained. Based on the first identification result, determine the first similarity; When the first similarity is greater than the preset similarity threshold, the sidewall image to be identified and the target tunnel sidewall image are input into the pre-trained target image recognition model for recognition, and a second recognition result is obtained. Based on the second identification result, determine the second similarity; When the second similarity is less than or equal to the preset spraying similarity threshold, the spraying completion status of multiple sidewall sub-regions is determined as spraying completion; When the second similarity is greater than the preset spraying similarity threshold, the spraying completion status of multiple sidewall sub-regions is determined to be spraying incomplete.

[0128] In some embodiments, the generating unit 403 is further configured to: Based on the initial tunnel sidewall images, the smoothness of the wall surface corresponding to each sidewall sub-region is identified; Based on the smoothness of the wall surface in each area, determine the spraying pressure parameters for each sidewall sub-area; Determine the region coordinate information corresponding to each sidewall sub-region; By combining the area coordinate information and spraying pressure parameters corresponding to each sidewall sub-region, multiple area spraying control parameters are generated.

[0129] In some embodiments, the state detection unit 405 is further configured to: The target tunnel sidewall image and the initial tunnel sidewall image are input into the pre-trained target image recognition model for recognition, and the target recognition result is obtained. The point cloud information of each side wall sub-region is obtained by LiDAR, and the smoothness of the wall surface of each side wall sub-region is determined based on the point cloud information of each region. When the target similarity corresponding to the target recognition result is less than or equal to the preset similarity threshold, and the smoothness of the area wall is greater than or equal to the preset smoothness threshold, the spraying completion status corresponding to multiple side wall sub-areas is determined as spraying completion. When the comparison similarity is greater than the preset similarity threshold, or the smoothness of the area wall is less than the preset smoothness threshold, the spraying completion status of multiple side wall sub-areas is determined to be spraying incomplete.

[0130] As described above, this application can determine the remaining path length of the target tunnel using a shotcrete robot, thereby determining the shotcrete operation progress of the target tunnel. Through image recognition, it identifies multiple sub-regions of the sidewall that need shotcreting based on the initial tunnel sidewall images collected at each target location, generating multiple area spraying control parameters for spraying control. Furthermore, it identifies the shotcrete completion status based on the collected target tunnel sidewall images. When the shotcreting of the tunnel sidewall area at the target location is completed, it continues to the next target location to perform the shotcrete operation until completion. Thus, it eliminates the need for manual grouting or manual control of the remote-controlled robot, instead using a shotcrete robot to perform adaptive shotcrete operations on the target tunnel. This provides flexibility and convenience, improves the efficiency of shotcrete operations on the target tunnel, reduces operation time, and increases the shotcrete operation progress of the target tunnel.

[0131] The specific implementation of each of the above units can be found in the previous embodiments, and will not be repeated here.

[0132] Figure 8To implement the structural block diagram of a terminal in this embodiment of the application, the terminal 110 includes: a radio frequency (RF) circuit 510, a memory 515, an input unit 530, a display unit 540, a sensor 550, an audio circuit 560, a wireless fidelity (WiFi) module 570, a processor 580, and a power supply 590, among other components. Those skilled in the art will understand that the terminal 110 structure shown in the figures does not constitute a limitation on a mobile phone or computer, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0133] The RF circuit 510 can be used to receive and transmit signals during information transmission or calls. In particular, it receives downlink information from the base station and processes it with the processor 580; in addition, it transmits uplink data to the base station.

[0134] The memory 515 can be used to store software programs and modules. The processor 580 executes various functional applications and data processing of the terminal by running the software programs and modules stored in the memory 515.

[0135] The input unit 530 can be used to receive input numeric or character information, and to generate key signal inputs related to the terminal's settings and function control. Specifically, the input unit 530 may include a touch panel 531 and other input devices 532.

[0136] The display unit 540 can be used to display input or provided information, as well as various menus of the terminal. The display unit 540 may include a display panel 541.

[0137] Audio circuit 560, speaker 561, and microphone 562 provide an audio interface.

[0138] In this embodiment, the processor 580 included in the terminal 110 can execute the control method of the shotcrete robot in the previous embodiment.

[0139] The terminal 110 in this application embodiment includes, but is not limited to, mobile phones, computers, intelligent voice interaction devices, smart home appliances, vehicle terminals, aircraft, robots, etc. This invention embodiment can be applied to various scenarios, including but not limited to cloud technology, artificial intelligence, smart transportation, and assisted driving.

[0140] This application also provides a computer-readable storage medium for storing program code for executing the control methods of the shotcrete robot in the foregoing embodiments.

[0141] This application also provides a computer program product, which includes a computer program. The processor of a computer device reads and executes the computer program, causing the computer device to perform the control method for the shotcrete robot described above.

[0142] Furthermore, the terms “comprising” and “including”, and any variations thereof, are intended to cover non-exclusive inclusion, such that a process, method, apparatus, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are expressly listed, but may include other steps or units that are not expressly listed or that are inherent to such process, method, product or device.

[0143] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0144] It should be understood that in the description of the embodiments of this application, "multiple" means two or more, "greater than", "less than", "exceeding" etc. are understood to exclude the number itself, and "above", "below", "within" etc. are understood to include the number itself.

[0145] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, or indirect coupling or communication connection between apparatuses or units, and may be electrical, mechanical, or other forms.

[0146] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0147] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0148] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0149] It should also be understood that the various implementation methods provided in this application can be combined arbitrarily to achieve different technical effects.

[0150] In the embodiments of this application, the terms "module" or "unit" refer to a computer program or part of a computer program that has a predetermined function and works with other related parts to achieve a predetermined goal, and can be implemented wholly or partially using software, hardware (such as processing circuitry or memory), or a combination thereof. Similarly, a processor (or multiple processors or memory) can be used to implement one or more modules or units. Furthermore, each module or unit can be part of an overall module or unit that includes the functionality of that module or unit.

[0151] The above is a detailed description of the embodiments of this application. However, this application is not limited to the above embodiments. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of this application. All such equivalent modifications or substitutions are included within the scope defined by the claims of this application.

Claims

1. A control method for a shotcrete robot, characterized in that, Applied to a shotcrete robot, the shotcrete robot includes a shotcrete assembly and a motor assembly, including: Determine the remaining path length corresponding to the target tunnel; When the remaining path length is greater than a preset length threshold, the motorized component is controlled to travel to the next target location point in the target tunnel, and the initial tunnel sidewall image of the surrounding tunnel sidewall of the target location point is acquired. Based on the initial tunnel sidewall image, multiple sidewall sub-regions corresponding to the target location point are determined, and multiple area spraying control parameters are generated based on the multiple sidewall sub-regions. According to the multiple area spraying control parameters, the spraying assembly is controlled to spray the multiple sidewall sub-areas in sequence, and the images of the surrounding roadway sidewalls of the target location point after spraying are collected. Based on the target tunnel sidewall image, determine the shotcreting completion status of the multiple sidewall sub-regions; When the shotcreting completion status is "shotcreting complete", return to the step of determining the remaining path length corresponding to the target tunnel, and stop the shotcreting operation on the target tunnel when the corresponding remaining path length is less than or equal to the preset length threshold.

2. The method according to claim 1, characterized in that, The step of determining multiple sidewall sub-regions corresponding to the target location point based on the initial tunnel sidewall image includes: The target tunnel sidewall region is determined based on the initial tunnel sidewall image, and the target total area corresponding to the target tunnel sidewall region is determined. From the list of preset division coefficients, query the target division coefficient corresponding to the target total area. The list of preset division coefficients includes multiple different preset total areas and the preset division coefficient corresponding to each preset total area. The target roadway sidewall region is divided according to the target division coefficient to obtain multiple sidewall sub-regions.

3. The method according to claim 1, characterized in that, The step of determining the shotcreting completion status of the multiple sidewall sub-regions based on the target tunnel sidewall image includes: The target tunnel sidewall image is identified using a pre-trained target image recognition model to obtain the recognition result. The pre-trained target image recognition model is obtained by minimizing the difference between the predicted recognition result output by the sample target alleyway sidewall image and the sample recognition result. Based on the identification results, the shotcreting completion status corresponding to the multiple sidewall sub-regions is determined.

4. The method according to claim 3, characterized in that, The shotcreting completion status is shotcreting incomplete. After determining the shotcreting completion status corresponding to the multiple sidewall sub-regions based on the identification result, the method further includes: Identify at least one target similarity region between the target tunnel sidewall image and the initial tunnel sidewall image; Based on the target coordinate information of the target similar area, generate the corresponding target area spraying control parameters; According to the target area spraying control parameters, the spraying assembly is controlled to spray and repair the corresponding target-similar area on the sidewall surrounding the target location point; Collect images of the sidewalls of the surrounding roadway at the target location after shotcrete repair; Based on the image of the sidewall to be identified, the shotcrete completion status of the multiple sidewall sub-regions is reassessed.

5. The method according to claim 4, characterized in that, The step of re-evaluating the shotcreting completion status corresponding to the multiple sidewall sub-regions based on the sidewall image to be identified includes: The image of the side wall to be identified is input into a pre-trained target image recognition model for recognition, and a first recognition result is obtained. When it is determined that spraying is required based on the first identification result, a first similarity is determined between the sidewall image to be identified and the initial tunnel sidewall image. When the first similarity is greater than a preset similarity threshold, a second similarity is determined between the sidewall image to be identified and the target alleyway sidewall image. When the second similarity is less than or equal to the preset spraying similarity threshold, the spraying completion status corresponding to the plurality of sidewall sub-regions is determined to be spraying completion; When the second similarity is greater than the preset spraying similarity threshold, the spraying completion status of the multiple sidewall sub-regions is determined to be spraying incomplete.

6. The method according to claim 1, characterized in that, The generation of multiple area spraying control parameters based on the multiple sidewall sub-regions includes: Based on the initial tunnel sidewall image, identify the surface smoothness of the corresponding area wall for each sidewall sub-region; Based on the smoothness of the wall surface in each area, determine the spraying pressure parameters for each sidewall sub-area; Determine the region coordinate information corresponding to each sidewall sub-region; By combining the region coordinate information and the spraying pressure parameters corresponding to each sidewall sub-region, multiple region spraying control parameters are generated.

7. The method according to claim 1, characterized in that, The step of determining the shotcreting completion status of the multiple sidewall sub-regions based on the target tunnel sidewall image includes: The target tunnel sidewall image is input into a pre-trained target image recognition model for recognition, and the target recognition result is obtained. The depth information of the depth camera is used to obtain the point cloud information of each side wall sub-region, and the smoothness of the wall surface of each side wall sub-region is determined based on the point cloud information of each region. When it is determined that spraying is not required based on the target recognition result, and the smoothness of the wall surface in the area is greater than or equal to a preset smoothness threshold, the spraying completion status corresponding to the multiple side wall sub-areas is determined to be spraying completion. When it is determined that spraying is required based on the target recognition result, or when the smoothness of the wall surface in the area is less than the preset smoothness threshold, the spraying completion status of the multiple side wall sub-areas is determined to be spraying incomplete.

8. A control device for a shotcrete robot, characterized in that, Applied to a shotcrete robot, the shotcrete robot includes a shotcrete assembly and a motor assembly, including: The determination unit is used to determine the remaining path length corresponding to the target roadway; The first control unit is used to control the motorized component to travel to the next target location point in the target tunnel when the remaining path length is greater than a preset length threshold, and to acquire the initial tunnel sidewall image of the surrounding tunnel sidewall of the target location point. The generation unit is used to determine multiple sidewall sub-regions corresponding to the target location point based on the initial tunnel sidewall image, and to generate multiple area spraying control parameters based on the multiple sidewall sub-regions. The second control unit controls the spraying assembly to spray the multiple sidewall sub-regions sequentially according to the multiple area spraying control parameters, and acquires images of the surrounding roadway sidewalls of the target location point after spraying. The status detection unit is used to determine the shotcreting completion status of the multiple sidewall sub-regions based on the target tunnel sidewall image. The repeat execution unit is used to return to the step of determining the remaining path length corresponding to the target tunnel when the shotcreting completion status is shotcreting complete, until the corresponding remaining path length is less than or equal to the preset length threshold and then stop the shotcreting operation on the target tunnel.

9. A computer device, characterized in that, The computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the control method of the shotcrete robot according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a plurality of instructions adapted for loading by a processor to execute the control method of the shotcrete robot according to any one of claims 1 to 7.

Images