A method and system for path planning of a water gauge-based negative pressure wall-climbing robot

Abstract

This invention relates to the field of draft survey technology, and discloses a path planning method and system for a negative pressure wall-climbing robot for draft surveying. The method includes: real-time acquisition of point cloud data and spectral reflectance data of the outer wall surface of a ship's cabin; dividing the outer surface of the cabin into several grid cells; calculating the adsorption stability score of each grid cell; generating a global navigation path based on the adsorption stability score and the target position using an improved fast random tree algorithm; dynamically allocating the adsorption force of the negative pressure wall-climbing robot based on the adsorption stability score of each grid cell, driving the negative pressure wall-climbing robot to move along the global navigation path; when the negative pressure wall-climbing robot reaches the target position along the global navigation path, photographing the draft coordinates of the outer wall surface of the cabin to obtain draft surveying data. This solution can enhance the adhesion ability of the wall-climbing robot on surfaces with varying curvature and contaminated surfaces, reduce the risk of the wall-climbing robot falling off, and improve the automation level and reliability of draft surveying operations.

Description

Technical Field

[0001] This invention belongs to the field of water level gauge weighing technology, and particularly relates to a path planning method and system for a water level gauge weighing negative pressure wall-climbing robot. Background Technology

[0002] Draft surveying is a crucial measurement method for cargo loading and unloading on ships, determining the cargo load by observing the draft marks on the hull. Traditional methods rely on manual climbing of the ship's side or suspended observation in a basket, which suffers from problems such as falls from heights and large reading errors. Solutions relying on unmanned surface vessels (USVs) or drones to remotely photograph draft readings are costly and highly susceptible to weather conditions, making operation impossible in severe weather conditions such as strong winds and heavy rain. While wall-climbing robots have emerged in recent years to replace manual labor, they still face technical bottlenecks such as adhesion failure and positioning drift in complex environments like the curved surfaces, oil stains, and weld seams of ship hulls.

[0003] Currently, existing wall-climbing robots rely on geometric point cloud data for path planning, which cannot identify chemical contaminants such as oil and sewage stains. When the robot passes through an oily area, the sealing ring fails, causing it to detach instantly. Although the problem can be temporarily alleviated by increasing the adsorption force, energy consumption increases dramatically and the risk of detachment cannot be completely eliminated. Most solutions separate path planning from adsorption control. After the planning module outputs a fixed path, the adsorption system maintains constant pressure throughout. When crossing ship cabin welds or curved surfaces, the failure to dynamically distribute the adsorption force according to the local curvature leads to leakage in the edge cavities, causing a chain reaction of failures.

[0004] Therefore, there is an urgent need to develop a path planning method and system for a negative pressure wall-climbing robot for water level gauge weighing, which can enhance the robot's adhesion to surfaces with varying curvature and contaminated surfaces, reduce the risk of the robot falling off, and improve the automation level and reliability of water level gauge weighing operations. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a path planning method and system for a negative pressure wall-climbing robot for water level gauge weighing. This method enhances the robot's adhesion to surfaces with varying curvature and contaminated surfaces, reduces the risk of the robot falling off, and improves the automation level and reliability of water level gauge weighing operations.

[0006] This invention provides a path planning method for a water level gauge-based negative pressure wall-climbing robot, the method comprising the following steps:

[0007] S1. The negative pressure wall-climbing robot collects point cloud data and spectral reflectance data of the outer wall surface of the cabin in real time, divides the outer wall surface of the cabin into several grid units, and calculates the adsorption stability score of each grid unit.

[0008] S2. Based on the adsorption stability score of each grid cell and the target location, an improved method is adopted. The algorithm generates a global navigation path;

[0009] S3. Based on the adsorption stability score of each grid unit, dynamically allocate the adsorption force of the negative pressure wall-climbing robot and drive the negative pressure wall-climbing robot to move along the global navigation path.

[0010] S4. When the negative pressure wall-climbing robot reaches the target position along the global navigation path, it takes pictures of the water gauge coordinates on the outer wall surface of the cabin to obtain water gauge weight data.

[0011] Furthermore, the negative pressure wall-climbing robot includes a binocular camera with an integrated multispectral analysis module and a nine-grid adsorption module. Each adsorption module is equipped with a vacuum pressure sensor and a solenoid valve in its cavity.

[0012] Furthermore, in S1, the negative pressure wall-climbing robot collects point cloud data and spectral reflectance data of the outer surface of the cabin in real time, divides the outer surface of the cabin into several grid cells, and calculates the adsorption stability score of each grid cell, including:

[0013] S11, the negative pressure wall-climbing robot scans the outer wall of the cabin with a binocular camera to generate three-dimensional point cloud data;

[0014] S12. Calculate the curvature distribution and protrusion line position of the outer wall surface of the cabin based on the three-dimensional point cloud data. Mark the grid cells with curvature below the threshold as adsorption stable regions and the grid cells with curvature above the threshold and the grid cells within the preset range of the protrusion line as adsorption unstable regions.

[0015] S13. Based on the spectral reflectance data, the grid cells with reflectance of the first wavelength less than the first preset value and the grid cells with reflectance of the second wavelength less than the second preset value are marked as contaminated areas.

[0016] S14. Calculate the geometric stability score of each grid cell based on the curvature of each grid cell.

[0017] S15. Define the pollution score of each grid cell based on whether it belongs to a polluted area;

[0018] S16. Calculate the adsorption stability score of each grid cell based on the geometric stability score and contamination score of each grid cell.

[0019] Furthermore, in S2, based on the adsorption stability score of each grid cell and the target location, an improved method is adopted. The algorithm generates a global navigation path, including:

[0020] S21. Calculate the navigation cost of each grid cell based on the stability score of each grid cell to obtain the navigation cost matrix;

[0021] S22. Using the target location as the endpoint and the center point of each grid cell as a candidate path node, an improved method is adopted. The algorithm generates a global navigation path;

[0022] Among them, improvements The algorithm includes The random sampling was changed to cost-sensitive sampling.

[0023] Furthermore, the calculation formula for cost-sensitive sampling is as follows:

[0024] ;

[0025] Where P(i,j) represents the sampling probability of the grid cell in the i-th row and j-th column, and C ij This represents the navigation cost of the grid cell in the i-th row and j-th column. This indicates the elimination of the zero constant.

[0026] Furthermore, in S22, an improved method is adopted. The algorithm also includes the following when generating global navigation paths:

[0027] Calculate the coverage area of ​​the nine-square grid adsorption module of the negative pressure wall-climbing robot;

[0028] Each candidate path node is used as the center point of the coverage area. It is determined whether there is a polluted area within the coverage area. If so, the candidate path node is deleted; otherwise, the candidate path node is retained.

[0029] Furthermore, in S3, based on the adsorption stability score of each grid cell, the adsorption force of the negative pressure wall-climbing robot is dynamically allocated, driving the negative pressure wall-climbing robot to move along the global navigation path, including:

[0030] S31. Construct a mapping relationship between adsorption stability score and required adsorption force, and calculate the target pressure difference based on the adsorption stability score of the grid unit corresponding to the cavity of each adsorption module of the negative pressure climbing robot.

[0031] S32. Obtain the current pressure difference of the cavities of each adsorption module of the negative pressure wall-climbing robot through a vacuum pressure sensor;

[0032] S33. A PID controller is used to adjust the valve opening of the solenoid valve of each adsorption module in real time according to the target pressure difference and the current pressure difference.

[0033] This invention also provides a path planning system for a water level gauge negative pressure wall-climbing robot, used to execute the above-described path planning method for a water level gauge negative pressure wall-climbing robot. The system includes the following modules:

[0034] The data acquisition module is used by the negative pressure wall-climbing robot to collect point cloud data and spectral reflectance data of the outer wall surface of the cabin in real time, divide the outer surface of the cabin into several grid units, and calculate the adsorption stability score of each grid unit.

[0035] The path generation module, connected to the data acquisition module, is used to determine the adsorption stability score and target location based on each grid cell, employing an improved... The algorithm generates a global navigation path;

[0036] The navigation module, connected to the path generation module, is used to dynamically allocate the adsorption force of the negative pressure wall-climbing robot based on the adsorption stability score of each grid cell, and drive the negative pressure wall-climbing robot to move along the global navigation path.

[0037] The output module, connected to the navigation module, is used to photograph the water gauge coordinates on the outer wall surface of the cabin when the negative pressure wall-climbing robot reaches the target position along the global navigation path, and obtain water gauge weight data.

[0038] The embodiments of the present invention have the following technical effects:

[0039] This invention captures surface curvature distribution using binocular vision and simultaneously identifies oil and water stains using multispectral analysis. It transforms contamination risk and surface smoothness into an adsorption stability score, overcoming the limitations of traditional path planning that relies solely on geometric features, thus improving the safety of path planning. The algorithm actively avoids low-stability and contaminated areas during the planning phase by using a grid navigation cost matrix and cost-sensitive sampling. It also introduces a nine-square grid coverage area verification to ensure that the robot body is completely free from the contaminated area. Based on the adsorption stability score, the adsorption force of each cavity in the nine-square grid is adjusted in real time. The high-stability area operates in an energy-saving manner, and the low-stability area is compensated by pressure boosting. This enhances the robot's adhesion ability on surfaces with varying curvature and contaminated surfaces, and solves the adsorption failure problem caused by contamination. Attached Figure Description

[0040] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0041] Figure 1 This is a flowchart of a path planning method for a negative pressure wall-climbing robot based on a water level gauge, provided in an embodiment of the present invention.

[0042] Figure 2 This is a structural schematic diagram of a water level gauge negative pressure wall-climbing robot provided in an embodiment of the present invention;

[0043] Figure 3 This is a schematic diagram of the path planning system for a negative pressure wall-climbing robot based on a water level gauge, provided in an embodiment of the present invention. Detailed Implementation

[0044] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0045] This invention provides a path planning method for a water level gauge-based negative pressure wall-climbing robot. Figure 1 This is a flowchart of a path planning method for a water level gauge-based negative pressure wall-climbing robot provided in an embodiment of the present invention. See also... Figure 1 The method includes the following steps:

[0046] S1. The negative pressure wall-climbing robot collects point cloud data and spectral reflectance data of the outer wall surface of the cabin in real time, divides the outer wall surface of the cabin into several grid units, and calculates the adsorption stability score of each grid unit.

[0047] In some embodiments, Figure 2 This is a structural schematic diagram of a water level gauge negative pressure wall-climbing robot provided in an embodiment of the present invention. See also: Figure 2 The negative pressure wall-climbing robot includes a binocular camera with an integrated multispectral analysis module and a nine-grid adsorption module. Each adsorption module is equipped with a vacuum pressure sensor and a solenoid valve in its cavity.

[0048] In the system design of the draft survey negative pressure wall-climbing robot, a binocular camera with integrated multispectral analysis function and a nine-grid adsorption module are used to achieve effective detection and stable adsorption on the hull surface. The binocular camera assembly not only provides stereoscopic vision information to generate a 3D point cloud map of the target area, but also integrates sensors sensitive to specific wavelengths to detect changes in the reflectivity of the hull surface. Through this dual capability, the device can accurately identify the location of structural features such as welds and reinforcing ribs, as well as contaminated areas with oil or water stains during scanning. Furthermore, each adsorption module is equipped with a vacuum pressure sensor and a solenoid valve, allowing for dynamic adjustment of the operating parameters of each adsorption module to adapt to different surface conditions as the negative pressure wall-climbing robot begins to move. The nine-grid adsorption module design aims to improve the robot's adaptability and stability in complex surface environments. This design allows for a wider adsorption area coverage and flexible handling of various obstacles or irregular shapes that may appear on the hull surface. Each adsorption module is connected to an air extraction pipe, and a fan generates the necessary negative pressure for each adsorption module. In addition, steering wheels are set on both sides of the negative pressure wall-climbing robot to enable it to smoothly complete directional changes.

[0049] In some embodiments, S1 includes the following sub-steps:

[0050] S11, the negative pressure wall-climbing robot scans the outer wall of the cabin with a binocular camera to generate three-dimensional point cloud data.

[0051] S12. Calculate the curvature distribution and protrusion line position of the outer wall surface of the cabin based on the three-dimensional point cloud data. Mark the grid cells with curvature below the threshold as adsorption stable regions, and mark the grid cells with curvature above the threshold and grid cells within the preset range of the protrusion line as adsorption unstable regions.

[0052] Specifically, point cloud data covers surface geometry features, including curvature (curvature variations) and protruding lines (such as welds or joints), which are factors that affect adsorption stability.

[0053] S13. Based on the spectral reflectance data, the grid cells with reflectance of the first wavelength less than the first preset value and the grid cells with reflectance of the second wavelength less than the second preset value are marked as contaminated areas.

[0054] Specifically, water stains cause a decrease in surface reflectivity, while oil stains exhibit specific spectral characteristics (e.g., engine oil shows a sudden increase in absorption in the 900-1000nm wavelength range). For example, by simultaneously collecting spectral data in the 400-1700nm wavelength range and identifying contaminants using a reflectivity model, water molecules exhibit a strong absorption peak at 1450nm, and surfaces covered by water stains significantly reduce reflectivity in this wavelength range. Hydrocarbons (the main component of oil stains) generate a CH bonding absorption peak at 1700nm, and oil film coverage forms a low-reflectivity area. Therefore, grid cells with reflectivity less than a first preset value (10%) at the first wavelength (1450nm) and grid cells with reflectivity less than a second preset value (8%) at the second wavelength (1700nm) can be marked as contaminated areas. The first and second preset values ​​can be determined experimentally by measuring the maximum reflectivity of a dry surface after being covered with a water or oil film.

[0055] S14. Calculate the geometric stability score of each grid cell based on the curvature of each grid cell.

[0056] In some embodiments, the geometric stability score is calculated using the following formula:

[0057] ;

[0058] Among them, S geo,ij k represents the geometric stability score of the grid cell in the i-th row and j-th column. ij This represents the curvature of the grid cell in the i-th row and j-th column. This score is negatively correlated with local curvature; that is, the smaller the curvature and the flatter the surface, the higher the score.

[0059] S15. Define the pollution score for each grid cell based on whether it belongs to a polluted area.

[0060] In some embodiments, the contamination score of each grid cell is defined as follows:

[0061] If it belongs to a polluted area, the pollution score of this grid cell is 1;

[0062] If a grid cell does not belong to a polluted area, its pollution score is 0.

[0063] S16. Calculate the adsorption stability score of each grid cell based on the geometric stability score and contamination score of each grid cell.

[0064] In some embodiments, the adsorption stability score is calculated using the following formula:

[0065] ;

[0066] Among them, H ij R represents the adsorption stability score of the grid cell in the i-th row and j-th column, a represents the influence weight of the geometric stability score, b represents the influence weight of the contamination score, and R poll,ij This represents the pollution score of the grid cell in the i-th row and j-th column. For example, the influence weights a and b can be set according to actual conditions; in this embodiment, a = 0.4 and b = 0.6.

[0067] S2. Based on the adsorption stability score of each grid cell and the target location, an improved method is adopted. The algorithm generates a global navigation path.

[0068] In some embodiments, S2 specifically includes the following sub-steps:

[0069] S21. Calculate the navigation cost of each grid cell based on the stability score of each grid cell to obtain the navigation cost matrix.

[0070] In generating a globally executable navigation path for the robot, the system first converts the previously obtained adsorption stability scores of each grid cell into navigation costs during the path search process. This conversion mechanism ensures that areas with poor stability are assigned higher passage costs, thus being prioritized for avoidance during the path planning phase. The establishment of the navigation cost matrix allows for a quantitative expression of the passage difficulty of the entire working area, providing a clear optimization objective for subsequent algorithms.

[0071] In some embodiments, the navigation cost of each grid cell is expressed as follows:

[0072] ;

[0073] Among them, C ijThis represents the navigation cost of the grid cell in the i-th row and j-th column. The navigation cost value and the threshold for the adsorption stability score in the above expression can be adaptively adjusted according to the actual situation, and are not limited here.

[0074] S22. Using the target location as the endpoint and the center point of each grid cell as a candidate path node, an improved method is adopted. The algorithm generates a global navigation path.

[0075] In this approach, the center point of each grid cell is used as a candidate path node, and the navigation cost of each grid cell is used as the navigation cost of that candidate path node.

[0076] Among them, improvements The algorithm includes The random sampling is changed to cost-sensitive sampling. The calculation formula for cost-sensitive sampling is as follows:

[0077] ;

[0078] Where P(i,j) represents the sampling probability of the grid cell in the i-th row and j-th column, and C ij This represents the navigation cost of the grid cell in the i-th row and j-th column. This indicates the elimination of the zero constant.

[0079] This algorithm makes key optimizations to the traditional fast expansion random tree algorithm by changing the original completely random sampling strategy to cost-sensitive sampling. This means that in each expansion step, the algorithm no longer randomly selects new nodes uniformly from the entire space, but dynamically adjusts the probability of selection based on the navigation cost of each grid cell. This mechanism guides the search process to focus more on low-cost, high-stability areas, significantly improving the efficiency and safety of path generation.

[0080] Areas with lower navigation costs have a higher sampling probability, making the randomly generated nodes more likely to appear on suitable surfaces. This mechanism effectively avoids the large amount of invalid expansion that may occur with traditional random sampling, reducing the waste of computational resources. During the gradual growth of the path tree, the system continuously evaluates the feasibility of connections between new and existing nodes and continuously optimizes the total cost of the generated paths through rewiring strategies. The entire planning process considers not only connectivity from the starting point to the ending point but also the overall safety and stability of the path. By introducing a cost-sensitive mechanism, the algorithm can quickly converge to an optimal or near-optimal path that balances passage efficiency and adsorption reliability in complex ship surface environments, providing high-quality guidance commands for the robot's subsequent autonomous movement.

[0081] Furthermore, adopt improved The algorithm also includes the following when generating global navigation paths:

[0082] Calculate the coverage area of ​​the nine-square grid adsorption module of the negative pressure wall-climbing robot;

[0083] Each candidate path node is used as the center point of the coverage area to determine whether there is a contaminated area within the coverage area.

[0084] If yes, delete the candidate path node; otherwise, keep the candidate path node.

[0085] In improvement When expanding the path tree, the center point of each grid cell considered as a candidate path node is regarded as a potential center location of the nine-square grid adsorption module. It is necessary to evaluate whether all grid cells within the nine-square grid area centered on that point meet the safe passage conditions; that is, to focus on detecting whether there are any cells previously identified as contaminated areas within this coverage area. Since contaminants may compromise vacuum sealing, leading to localized adsorption failure, if any contaminated cell is found within the nine-square grid coverage area, that location is considered unable to provide stable and reliable attachment support for the robot. Therefore, this candidate node will be removed from the feasible path set to avoid subsequent paths passing through such locations with adsorption risks. Only when all grid cells within the entire nine-square grid coverage area are not contaminated areas and the overall stability score is at an acceptable level will the node be retained and participate in path connection and optimization.

[0086] This screening mechanism essentially integrates the robot's physical dimensions and adsorption characteristics into the constraints of path planning, achieving an upgrade in safety assessment from a "point" to a "surface" perspective. It considers not only the properties of individual grid cells but also the environmental compatibility of the robot's overall contact surface, preventing overall instability caused by local defects. Through this method, the robot can select a path truly suitable for the stable operation of its physical structure on complex ship surfaces, significantly improving the safety and reliability of movement on non-ideal surfaces such as oil stains and rust, and avoiding the risk of slippage or detachment due to insufficient adsorption.

[0087] S3. Based on the adsorption stability score of each grid unit, dynamically allocate the adsorption force of the negative pressure wall-climbing robot and drive the negative pressure wall-climbing robot to move along the global navigation path.

[0088] In some embodiments, S3 includes the following sub-steps:

[0089] S31. Construct a mapping relationship between adsorption stability score and required adsorption force, and calculate the target pressure difference based on the adsorption stability score of the grid unit corresponding to the cavity of each adsorption module of the negative pressure climbing robot.

[0090] When the robot moves to a certain position, the stability score of each unit covered by its nine-grid adsorption module is read in real time, and then the target pressure difference value that each adsorption chamber should reach is calculated. This target pressure difference reflects the vacuum strength required to maintain stable adsorption. Areas with lower scores correspond to higher target pressure differences to enhance adsorption reliability; conversely, in stable areas, the target pressure difference can be appropriately reduced to help with energy-saving operation.

[0091] S32. Obtain the current pressure difference of the cavities of each adsorption module of the negative pressure wall-climbing robot through a vacuum pressure sensor.

[0092] S33. A PID controller is used to adjust the valve opening of the solenoid valve of each adsorption module in real time according to the target pressure difference and the current pressure difference.

[0093] The expression is as follows:

[0094] ;

[0095] Among them, K vq Let e(t) represent the valve opening degree of the solenoid valve in the q-th adsorption module. q K represents the deviation between the current pressure difference and the target pressure difference of the q-th adsorption module. p K represents the proportionality coefficient. i K represents the integral coefficient. d This represents the differential coefficient, which can be set according to the actual situation. For example, in this embodiment, K... p =0.8, K i =0.05, K d =0.3.

[0096] The control unit employs a PID control algorithm to process the deviation between the target pressure difference and the actual pressure difference. Based on the magnitude and trend of the deviation, the valve openings of the solenoid valves connecting each adsorption chamber are adjusted in real time. If the current pressure difference is lower than the target value, the controller increases the solenoid valve opening to increase the pumping rate and accelerate vacuum establishment; if the pressure difference approaches or reaches the set value, the opening is reduced to maintain a stable negative pressure; when the pressure difference is too high, the valves are appropriately closed to avoid excessive energy consumption or unnecessary stress on the surface. Since all adsorption modules share a single fan to generate negative pressure, the independent adjustment capability of the solenoid valves allows for differentiated control of the vacuum level in each chamber, thereby achieving flexible distribution of adsorption force on the robot's bottom surface. This closed-loop control mechanism ensures that the robot maintains overall adhesion stability even when traversing areas of curvature change, weld edges, or localized contamination.

[0097] S4. When the negative pressure wall-climbing robot reaches the target position along the global navigation path, it takes pictures of the water gauge coordinates on the outer wall surface of the cabin to obtain water gauge weight data.

[0098] When the negative pressure wall-climbing robot reaches the target position along the global navigation path, it rotates its binocular camera to capture the water gauge coordinates on the outer wall surface of the cabin, thus obtaining water gauge weight data.

[0099] This invention also provides a path planning system for a water level gauge negative pressure wall-climbing robot, used to execute the aforementioned path planning method for a water level gauge negative pressure wall-climbing robot. Figure 3 This is a schematic diagram of a path planning system for a negative pressure wall-climbing robot based on a water level gauge, provided in an embodiment of the present invention. (See attached diagram.) Figure 3 The system includes the following modules:

[0100] The data acquisition module is used by the negative pressure wall-climbing robot to collect point cloud data and spectral reflectance data of the outer wall surface of the cabin in real time, divide the outer surface of the cabin into several grid units, and calculate the adsorption stability score of each grid unit.

[0101] The path generation module, connected to the data acquisition module, is used to determine the adsorption stability score and target location based on each grid cell, employing an improved... The algorithm generates a global navigation path;

[0102] The navigation module, connected to the path generation module, is used to dynamically allocate the adsorption force of the negative pressure wall-climbing robot based on the adsorption stability score of each grid cell, and drive the negative pressure wall-climbing robot to move along the global navigation path.

[0103] The output module, connected to the navigation module, is used to photograph the water gauge coordinates on the outer wall surface of the cabin when the negative pressure wall-climbing robot reaches the target position along the global navigation path, and obtain water gauge weight data.

[0104] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the technical solutions of the embodiments of the present invention.

Claims

1. A path planning method for a water level gauge-based negative pressure wall-climbing robot, characterized in that, The method includes the following steps: S1. The negative pressure wall-climbing robot collects point cloud data and spectral reflectance data of the outer wall surface of the cabin in real time, divides the outer surface of the cabin into several grid units, and calculates the adsorption stability score of each grid unit. Specifically, it includes: S11. The negative pressure wall-climbing robot scans the outer wall of the cabin with a binocular camera to generate three-dimensional point cloud data; S12. Calculate the curvature distribution and protrusion line position of the outer wall surface of the cabin based on the three-dimensional point cloud data. Mark grid cells with curvature below the threshold as adsorption stable regions and grid cells with curvature above the threshold and grid cells within the preset range of the protrusion line as adsorption unstable regions. S13. Based on the spectral reflectance data, the grid cells with reflectance of the first wavelength less than the first preset value and the grid cells with reflectance of the second wavelength less than the second preset value are marked as contaminated areas. S14. Calculate the geometric stability score of each grid cell based on the curvature of each grid cell. S15. Define the pollution score of each grid cell based on whether it belongs to a polluted area; S16. Calculate the adsorption stability score of each grid cell based on the geometric stability score and contamination score of each grid cell. S2. Based on the adsorption stability score and target position of each grid unit, an improved method is adopted. The algorithm generates a global navigation path; S3. Based on the adsorption stability score of each grid unit, dynamically allocate the adsorption force of the negative pressure wall-climbing robot and drive the negative pressure wall-climbing robot to move along the global navigation path; S4. When the negative pressure wall-climbing robot reaches the target position along the global navigation path, it takes a picture of the water gauge coordinates on the outer wall surface of the cabin to obtain water gauge weight data.

2. The path planning method for a water level gauge-based negative pressure wall-climbing robot according to claim 1, characterized in that, The negative pressure wall-climbing robot includes a binocular camera with an integrated multispectral analysis module and a nine-grid adsorption module. Each adsorption module is equipped with a vacuum pressure sensor and a solenoid valve in its cavity.

3. The path planning method for a water level gauge-based negative pressure wall-climbing robot according to claim 2, characterized in that, In step S2, based on the adsorption stability score and target position of each grid unit, an improved method is adopted. The algorithm generates a global navigation path, including: S21. Calculate the navigation cost of each grid cell based on the stability score of each grid cell to obtain the navigation cost matrix; S22. Using the target location as the endpoint and the center point of each grid cell as a candidate path node, an improved method is adopted. The algorithm generates a global navigation path; Among them, improvements The algorithm includes The random sampling was changed to cost-sensitive sampling.

4. The path planning method for a water level gauge-based negative pressure wall-climbing robot according to claim 3, characterized in that, The calculation formula for the cost-sensitive sampling is as follows: ； Where P(i,j) represents the sampling probability of the grid cell in the i-th row and j-th column, and C ij This represents the navigation cost of the grid cell in the i-th row and j-th column. This indicates the elimination of the zero constant.

5. The path planning method for a water level gauge-based negative pressure wall-climbing robot according to claim 3, characterized in that, In S22, an improved method is adopted. The algorithm also includes the following when generating global navigation paths: Calculate the coverage area of ​​the nine-square grid adsorption module of the negative pressure wall-climbing robot; Each candidate path node is used as the center point of the coverage area. It is determined whether there is a polluted area within the coverage area. If so, the candidate path node is deleted; otherwise, the candidate path node is retained.

6. The path planning method for a water level gauge-based negative pressure wall-climbing robot according to claim 2, characterized in that, In step S3, based on the adsorption stability score of each grid unit, the adsorption force of the negative pressure wall-climbing robot is dynamically allocated to drive the negative pressure wall-climbing robot to move along the global navigation path, including: S31. Construct a mapping relationship between adsorption stability score and required adsorption force, and calculate the target pressure difference based on the adsorption stability score of the grid unit corresponding to the cavity of each adsorption module of the negative pressure climbing robot. S32. Obtain the current pressure difference of the cavities of each adsorption module of the negative pressure wall-climbing robot through the vacuum pressure sensor; S33. A PID controller is used to adjust the valve opening of the solenoid valve of each adsorption module in real time according to the target pressure difference and the current pressure difference.

7. A path planning system for a water level gauge negative pressure wall-climbing robot, used to execute the path planning method for a water level gauge negative pressure wall-climbing robot according to any one of claims 1-6, characterized in that, The system includes the following modules: The data acquisition module is used by the negative pressure wall-climbing robot to collect point cloud data and spectral reflectance data of the outer wall surface of the cabin in real time, divide the outer surface of the cabin into several grid units, and calculate the adsorption stability score of each grid unit. The path generation module, connected to the data acquisition module, is used to perform improved [method / approach] based on the adsorption stability score and target location of each grid unit. The algorithm generates a global navigation path; The navigation module, connected to the path generation module, is used to dynamically allocate the adsorption force of the negative pressure wall-climbing robot based on the adsorption stability score of each grid unit, and drive the negative pressure wall-climbing robot to move along the global navigation path. The output module, connected to the navigation module, is used to photograph the water gauge coordinates on the outer wall surface of the cabin when the negative pressure wall-climbing robot reaches the target position along the global navigation path, and obtain water gauge weight data.

Images