Automatic height adjusting method of intelligent weeding machine based on terrain height perception

Abstract

The invention relates to the technical field of equipment adjustment, in particular to an intelligent weeding machine automatic height adjustment method based on terrain height perception, which comprises the following steps: acquiring depth data and terrain data of an operation land parcel through a multi-mode perception device; based on the acquired depth data and topographic data, reconstructing a three-dimensional model of the working plot by using a three-dimensional topographic reconstruction algorithm, and extracting regional topographic features; generating an optimal operation path by using a path planning algorithm; dynamically calculating the target height of the weeding execution mechanism; the cutter angle of the weeding mechanism is adjusted through an angle optimization algorithm so that the cutter angle can be attached to the terrain at the bottom of the gully; and executing target height and angle adjustment. Compared with a traditional weeding machine which only depends on a fixed height or a simple ground detection method, the weeding machine has the advantages that fine adjustment can be carried out according to topographic changes, the weeding efficiency and the operation quality are greatly improved, and missing weeding or excessive weeding is avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of equipment adjustment, and in particular to an automatic height adjustment method for an intelligent lawn mower based on terrain height perception. Background Art

[0002] With the advancement of agricultural modernization, intelligent weeding equipment has gradually become an important tool for farmland operations. Traditional weeders usually rely on manually set heights and angles to operate, and are difficult to adapt to complex terrain changes, especially in areas with large slopes or gullies. Weeds in these areas are often missed, resulting in unsatisfactory weeding effects. In order to address these problems, many researchers have tried to improve the operating accuracy and adaptability of weeders through automated control systems and sensor technology, but most of the existing technologies have the following shortcomings: Insufficient terrain perception: Traditional lawn mowers rely on simple sensors to measure ground height, but these sensors may have blind spots or errors in complex terrain (gullies, steep slopes) and cannot fully perceive the actual terrain changes of the land. This causes the lawn mower to be unable to accurately adjust the operating height and angle when encountering complex terrain, thereby reducing the operating effect.

[0003] Low efficiency of operation path planning: When planning the operation path of the lawn mower, the existing path planning algorithm usually only plans based on fixed operation parameters (fixed height, preset path), but ignores the actual terrain characteristics of the operation site. This simple path planning method causes the lawn mower to be inefficient during operation, especially in complex terrain areas, which often requires multiple adjustments back and forth, wasting a lot of time and energy.

[0004] The angle adjustment of the weeding tool is not flexible: In existing weeding equipment, the adjustment of the tool angle is often relatively simple, and lacks the ability to adapt to terrain changes in real time. For gullies or irregular terrain, the tool angle cannot be adjusted in real time to adapt to terrain changes, resulting in the tool not fitting the ground, causing low operating efficiency and difficulty in removing grass roots. Summary of the invention

[0005] The invention provides an automatic height adjustment method for an intelligent weeder based on terrain height perception.

[0006] An automatic height adjustment method for an intelligent lawn mower based on terrain height perception comprises the following steps: S1, data collection of the work site: Through the multi-modal sensing device, the depth data and terrain data of the work site are collected, including: The laser radar obtains the surface laser point cloud data of the working area; Ultrasonic sensors capture local gullies in the land; S2, Land parcel depth feature analysis and zoning modeling: Based on the collected depth data and terrain data, a 3D terrain reconstruction algorithm is used to reconstruct the 3D model of the work parcel and extract regional terrain features: Regional slope change trends; gully depth and gully distribution; S3, operation path planning: Combined with the regional terrain features extracted from the operation plot model, the path planning algorithm is used to generate the optimal operation path to ensure that the weeding execution mechanism enters the gully area first; S4, target height calculation: based on the optimal operation path, when the weeder is weeding according to the optimal operation path, the target height of the weeding actuator is dynamically calculated; S5, gully angle processing: In the gully area, the angle of the cutter of the weeding mechanism is adjusted through the angle optimization algorithm to make it fit the terrain at the bottom of the gully to ensure that the weeds are completely removed; S6, real-time adjustment of height and angle: The weeding actuator is driven by a multi-degree-of-freedom robotic arm to perform target height and angle adjustment.

[0007] Optionally, the laser radar scans the surface of the work site to obtain high-precision laser point cloud data, and captures the ground height and slope of the work site through multi-angle scanning and laser beam rotation. The ultrasonic sensor is used for local ultrasonic data acquisition, emitting ultrasonic pulses and receiving reflected waves in real time, and detecting ground mutation areas, which are gully areas (if the height data of a certain point changes significantly relative to the surrounding data, this can be achieved by setting a height mutation threshold. When the ultrasonic sensor detects the work site, a certain point suddenly exceeds the height mutation threshold, i.e., a mutation occurs); the collected ultrasonic data provides local gully information to supplement the blind spots that may appear in laser point cloud data in complex terrain areas (such as gullies and gullies).

[0008] Optionally, the S2 specifically includes: By fusing the collected laser point cloud data and ultrasonic sensor data, a complete three-dimensional model of the work site is generated using a three-dimensional terrain reconstruction algorithm. The laser point cloud data is represented as ,in, is the 3D coordinate of the i-th point in the point cloud, describing the overall 3D spatial structure of the work site. The ultrasonic sensor data is expressed as ,in is the local height data of the jth point measured by the ultrasonic sensor, is the 2D plane coordinate of the jth point, is the height value of the jth point, describing the local height data, which is used to supplement and correct the laser point cloud data during data fusion; The three-dimensional terrain reconstruction algorithm specifically includes: Delaunay triangulation: In the laser point cloud data, the collected point cloud data is triangulated by the DelaunayTriangulation algorithm, and the generated triangular mesh Each pair of adjacent points in the point cloud is connected to form a preliminary model; Surface fitting and interpolation: Using known point cloud data, calculate and supplement missing surface points through interpolation methods , x, y, z are the three-dimensional coordinates of the missing surface point p, x, y represent the plane coordinates of the surface point on the surface of the plot, and z represents the ground height of the surface point, so that the surface morphology of the entire plot is smooth and continuous. Spline interpolation can be used to obtain a smooth height function , describing the height of the land surface.

[0009] Optionally, the slope change trend of the area is calculated on the reconstructed three-dimensional model of the work plot, and the slope (i.e., the inclination of the ground) is determined by calculating the direction of the ground normal vector. Given any point , the slope S is calculated as follows: ; Among them, arctan is the inverse tangent function, and They are height functions The partial derivatives in the x and y directions are discretized using the gradient operator. By traversing the grid of the entire operation plot, the slope value of each grid cell is obtained and the slope change trend is calculated. Extracted through statistical regression analysis.

[0010] Optionally, the gully distribution is displayed by marking all gully areas in the three-dimensional model of the working plot; Based on the ground points in the three-dimensional model of the operation plot, the gully depth of each gully area is calculated, and the laser point cloud data middle, is the ground height at the corresponding position, and the relative depth of each gully area is calculated using the ground height difference: , where z1 is the average height of the adjacent non-gully area, and z2 is the height of the lowest point of the gully. By calculating in this way, the relative depth value of each position can be obtained.

[0011] Optionally, the operation path planning in S3 adopts The algorithm prioritizes the gully area in combination with the terrain characteristics and sets the gully area as a high weight to ensure that the path planning prioritizes the gully area as part of the operation path. When calculating the path, for each operation point, the cost function to the target point is calculated to adjust the impact of various factors on the path selection.

[0012] Optionally, the The algorithm specifically includes: S31, cost function definition: ensure that the weeding actuator enters the gully area first, design the cost function , the cost function reflects each grid point The difficulty level is: Slope cost: Areas with large slopes increase the difficulty of operations, so the slope cost is defined as: ,in is the slope weight coefficient, is the grid point The slope of the Gully cost: Gully areas have complex terrain and higher cost, defined as: ,in, is the weight coefficient of gully depth, is the grid point The depth of the gully at Total cost function Take the weighted sum of the two: ; S32, use The algorithm performs path search. The heuristic function of the algorithm Estimate the shortest distance from the current point to the target point, the cost function As the cost of the path, the goal is to find the path with the minimum total cost: ,in, is the actual cost from the starting point to the current grid point, the actual cost is the sum of the costs of all grid points from the starting point to the current grid point. The cost of each step comes from : , which means that each time you move from the starting point to a certain grid point , accumulate terrain cost along the path , For the heuristic estimate of the cost from the current grid point to the target grid point, choose the total cost The lowest path is taken as the optimal operation path.

[0013] Optionally, the target operating height of the weeding actuator in S4 Dynamically adjust according to the regional terrain characteristics of the current operation point: Flat area: If the current working grid point If the area is flat, the target working height is equal to the ground height. equal; Slope area: For areas with large slopes, Dynamically adjust the target working height to avoid collision with the ground or leaving the effective working range: ,in, is the slope adjustment factor; Gully Areas: For gully areas, the target working height should be relative to the lowest point of the gully Adjustments are made to ensure that the weeding actuator is adapted to the needs of deep trench operations: ,in, is the lowest point height of the current gully area, It is a preset safe working gap used to ensure the working distance of the weeding actuator.

[0014] Optionally, the S5 specifically includes: Based on the data provided by the ultrasonic sensor, the surface at the bottom of the gully is identified, and the inclination angle of the gully area is determined according to the gully depth information. The normal direction of the local surface at the bottom of the gully is set as , the direction is calculated as: , where z is at point The ground height of the location, and are the gradients of the point in the x and y directions respectively; By calculating the normal vector of the bottom of the gully And the original installation angle of the tool The angle between the two is used to determine the adjustment angle of the tool. Then, according to the direction of the normal vector, the target angle of the tool is calculated. : ,in, is the current direction vector of the tool, represents the dot product of vectors, and are the magnitudes of the vectors respectively.

[0015] Beneficial effects of the present invention: The present invention, by combining laser radar and ultrasonic sensors to collect high-precision depth data and terrain data of the working area, can sense terrain changes, especially complex areas such as gullies. This precise terrain perception enables the weeder to dynamically adjust its operating height and tool angle, especially in gully areas, to ensure that the weeding mechanism fits perfectly with the ground shape. Compared with traditional weeders that can only rely on fixed heights or simple ground detection methods, it can make fine adjustments based on terrain changes, greatly improve weeding efficiency and operation quality, and avoid missed or excessive weeding.

[0016] The present invention utilizes a cost function based on terrain features (such as slope, gully depth, etc.) to plan the optimal operating path, taking into account the terrain difficulty of different areas, and plans an optimal path for the lawn mower that is both efficient and can give priority to complex areas such as gullies. By minimizing the operating cost, the lawn mower can intelligently select the most suitable operating path, which not only improves the operating efficiency, but also ensures priority processing of difficult areas, avoiding the traditional path planning method's neglect of complex terrain.

[0017] The present invention adopts an angle optimization algorithm in the gully area to dynamically adjust the tool angle of the weeding mechanism based on the local terrain normal direction at the bottom of the gully, so that the weeder can accurately adjust the tool angle according to the specific terrain of the gully, thereby ensuring that the tool fits the bottom of the gully. This can not only improve the accuracy of weeding operations, but also effectively prevent the tool from getting stuck or damaged when processing low-lying or inclined areas, significantly improving the reliability of weeding operations and the ability to operate efficiently for a long time. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 A schematic diagram of a flow chart of an adjustment method according to an embodiment of the present invention; Figure 2 Schematic diagram of a cost function according to an embodiment of the present invention. DETAILED DESCRIPTION

[0020] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternatives to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0021] like Figure 1-Figure 2 As shown, a method for automatic height adjustment of an intelligent lawn mower based on terrain height perception includes the following steps: S1, data collection of the work site: Through the multi-modal sensing device, the depth data and terrain data of the work site are collected, including: The laser radar obtains the surface laser point cloud data of the working area; Ultrasonic sensors capture local gullies in the land; S2, Land parcel depth feature analysis and zoning modeling: Based on the collected depth data and terrain data, a 3D terrain reconstruction algorithm is used to reconstruct the 3D model of the work parcel and extract regional terrain features: Regional slope change trends; gully depth and gully distribution; S3, operation path planning: Combined with the regional terrain features extracted from the operation plot model, the path planning algorithm is used to generate the optimal operation path to ensure that the weeding execution mechanism enters the gully area first. Deep gully areas are often more difficult to cross. Prioritizing weeding operations in these areas can avoid repeatedly entering and exiting the same plot during the operation, thereby reducing interference with the operation path and improving overall efficiency. S4, target height calculation: based on the depth characteristics of each point in the optimal operation path, dynamically calculate the target height of the weeding actuator when the weeder is weeding according to the optimal operation path; S5, gully angle processing: In the gully area, the angle of the cutter of the weeding mechanism is adjusted through the angle optimization algorithm to make it fit the terrain at the bottom of the gully to ensure that the weeds are completely removed; S6, real-time adjustment of height and angle: The weeding actuator is driven by a multi-degree-of-freedom robotic arm to perform target height and angle adjustment.

[0022] The laser radar scans the surface of the working area to obtain high-precision laser point cloud data. Through multi-angle scanning and laser beam rotation, the ground height and slope of the working area are captured. The ultrasonic sensor is used for local ultrasonic data collection, emitting ultrasonic pulses and receiving reflected waves in real time to detect ground mutation areas. The mutation area is the gully area (if the height data of a certain point changes significantly relative to the surrounding data, it can be achieved by setting a height mutation threshold. When the ultrasonic sensor detects the working area, a certain point suddenly exceeds the height mutation threshold, that is, a mutation occurs); the collected ultrasonic data provides local gully information to supplement the blind spots that may appear in laser point cloud data in complex terrain areas (such as gullies and gullies).

[0023] S2 specifically includes: By fusing the collected laser point cloud data and ultrasonic sensor data, a complete three-dimensional model of the work site is generated using a three-dimensional terrain reconstruction algorithm. The laser point cloud data is represented as ,in, is the 3D coordinate of the i-th point in the point cloud, describing the overall 3D spatial structure of the work site. The ultrasonic sensor data is expressed as ,in is the local height data of the jth point measured by the ultrasonic sensor, is the 2D plane coordinate of the jth point, It is the height value of the jth point, describing the local height data, which is used to supplement and correct the laser point cloud data during data fusion. The ultrasonic data supplements and corrects the local details of the point cloud data, especially in areas where the point cloud data is sparse or has errors (such as gullies or sudden terrain). If there is a significant difference between the height value of the ultrasonic data and the height of the point cloud data, it can be replaced with the point cloud data to correct the height of the point; The 3D terrain reconstruction algorithm specifically includes: Delaunay triangulation: In the laser point cloud data, the collected point cloud data is triangulated by the DelaunayTriangulation algorithm, and the generated triangular mesh Each pair of adjacent points in the point cloud is connected to form a preliminary model; Surface fitting and interpolation: Using known point cloud data, calculate and supplement missing surface points through interpolation methods , x, y, z are the three-dimensional coordinates of the missing surface point p, x, y represent the plane coordinates of the surface point on the surface of the plot, and z represents the ground height of the surface point, so that the surface morphology of the entire plot is smooth and continuous. Spline interpolation can be used to obtain a smooth height function , describing the height of the land surface.

[0024] Regional slope change trend On the reconstructed 3D model of the work site, the slope (i.e. the inclination of the ground) is calculated, and the slope magnitude is determined by calculating the direction of the ground normal vector. , the slope S is calculated as follows: ; Among them, arctan is the inverse tangent function, and They are height functions The partial derivatives in the x and y directions are discretized using the gradient operator. By traversing the grid of the entire operation plot, the slope value of each grid cell is obtained and the slope change trend is calculated. Extracted through statistical regression analysis.

[0025] Assume the work area is a discrete grid , the slope of each grid point The calculation is as follows: , where the partial derivative of the slope is approximated by the difference method: ,in is the spatial resolution of the grid.

[0026] The gully distribution is displayed by marking all gully areas in the three-dimensional model of the working plot; Based on the ground points in the 3D model of the work area, the gully depth of each gully area is calculated, and the laser point cloud data middle, is the ground height at the corresponding position, and the relative depth of each gully area is calculated using the ground height difference: , where z1 is the average height of the adjacent non-gully area, and z2 is the height of the lowest point of the gully. By calculating in this way, the relative depth value of each position can be obtained.

[0027] The job path planning in S3 adopts The algorithm prioritizes the gully area in combination with the terrain characteristics and sets the gully area as a high weight to ensure that the path planning prioritizes the gully area as part of the operation path. When calculating the path, for each operation point, the cost function to the target point is calculated to adjust the impact of various factors on the path selection.

[0028] The algorithm specifically includes: S31, cost function definition: ensure that the weeding actuator enters the gully area first, design the cost function , the cost function reflects each grid point The difficulty level is: Slope cost: Areas with large slopes increase the difficulty of operations, so the slope cost is defined as: ,in is the slope weight coefficient, is the grid point The slope of the Gully cost: Gully areas have complex terrain and higher cost, defined as: ,in, is the weight coefficient of gully depth, is the grid point The depth of the gully at Total cost function Take the weighted sum of the two: ; Make areas with steeper slopes and deeper gullies have higher costs, thus guiding the path to prioritize entering these more difficult areas; S32, use The algorithm performs path search. The heuristic function of the algorithm Estimate the shortest distance from the current point to the target point, the cost function As the cost of the path, the goal is to find the path with the minimum total cost: ,in, is the actual cost from the starting point to the current grid point, the actual cost is the sum of the costs of all grid points from the starting point to the current grid point. The cost of each step comes from : , which means that each time you move from the starting point to a certain grid point , accumulate terrain cost along the path , For the heuristic estimate of the cost from the current point to the target grid point, select the total cost The lowest path is used as the optimal operation path, so as to ensure that the weeder enters the gully area first during operation. After completing the path planning, the weeder performs the operation according to the planned path, ensuring that the weeding execution mechanism enters the gully area first, especially the deep gully area. This priority strategy can significantly improve the efficiency of weeding operations and ensure the processing of complex areas.

[0029] Target operating height of weeding actuator in S4 Dynamically adjust according to the regional terrain characteristics of the current operation point: Flat area: If the current working grid point If the area is flat, the target working height is equal to the ground height. equal; Slope area: For areas with large slopes, Dynamically adjust the target working height to keep the weeding actuator at an appropriate height to avoid collision with the ground or leaving the effective working range: ,in, is the slope adjustment factor, which is set according to the actual situation; Gully Areas: For gully areas, the target working height should be relative to the lowest point of the gully Adjustments are made to ensure that the weeding actuator is adapted to the needs of deep trench operations: ,in, is the lowest point height of the current gully area, It is a preset safe working gap used to ensure the working distance of the weeding actuator.

[0030] S5 specifically includes: Based on the data provided by the ultrasonic sensor, the surface at the bottom of the gully is identified, and the inclination angle of the gully area is determined according to the gully depth information. The normal direction of the local surface at the bottom of the gully is set as , the direction is calculated as: , where z is at point The ground height of the location, and are the gradients of the point in the x and y directions respectively; By calculating the normal vector of the bottom of the gully And the original installation angle of the tool The angle between the two is used to determine the adjustment angle of the tool. Then, according to the direction of the normal vector, the target angle of the tool is calculated. : ,in, is the current direction vector of the tool, represents the dot product of vectors, and They are the modulus lengths of the vectors respectively. Through this angle calculation, the working angle of the tool can be adjusted to be consistent with the normal direction of the bottom of the gully.

[0031] Calculate the target angle of the cutter when the mower moves into the gully area The servo motor control system drives the angle adjustment of the tool to ensure that the tool always fits the terrain at the bottom of the gully to achieve precise weeding effect.

[0032] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.

[0033] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for automatic height adjustment of an intelligent lawn mower based on terrain height perception, characterized in that: The following steps are involved: S1, data collection of the work site: Through the multi-modal sensing device, the depth data and terrain data of the work site are collected, including: The laser radar obtains the surface laser point cloud data of the working area; Ultrasonic sensors capture local gullies in the land; S2, Land parcel depth feature analysis and zoning modeling: Based on the collected depth data and terrain data, a 3D terrain reconstruction algorithm is used to reconstruct the 3D model of the work parcel and extract regional terrain features: Regional slope change trends; gully depth and gully distribution; S3, operation path planning: Combined with the regional terrain features extracted from the operation plot model, the path planning algorithm is used to generate the optimal operation path to ensure that the weeding execution mechanism enters the gully area first; S4, target height calculation: based on the optimal operation path, when the weeder is weeding according to the optimal operation path, the target height of the weeding actuator is dynamically calculated; S5, gully angle processing: In the gully area, the angle of the cutter of the weeding mechanism is adjusted through the angle optimization algorithm to make it fit the terrain at the bottom of the gully; S6, real-time adjustment of height and angle: The weeding actuator is driven by a multi-degree-of-freedom robotic arm to perform target height and angle adjustment.

2. The method for automatic height adjustment of an intelligent lawn mower based on terrain height perception according to claim 1, characterized in that: The laser radar scans the surface of the work site to obtain laser point cloud data, and captures the ground height and slope of the work site through multi-angle scanning and laser beam rotation. The ultrasonic sensor is used for local ultrasonic data collection, emitting ultrasonic pulses and receiving reflected waves in real time to detect ground mutation areas, which are gully areas.

3. The method for automatic height adjustment of an intelligent lawn mower based on terrain height perception according to claim 1, characterized in that: The S2 specifically includes: By fusing the collected laser point cloud data and ultrasonic sensor data, a complete three-dimensional model of the work site is generated using a three-dimensional terrain reconstruction algorithm. The laser point cloud data is represented as ,in, is the 3D coordinate of the i-th point in the point cloud, describing the overall 3D spatial structure of the work site. The ultrasonic sensor data is expressed as ,in is the local height data of the jth point measured by the ultrasonic sensor, is the 2D plane coordinate of the jth point, is the height value of the jth point, describing the local height data, which is used to supplement and correct the laser point cloud data during data fusion; The three-dimensional terrain reconstruction algorithm specifically includes: Delaunay triangulation: In the laser point cloud data, the collected point cloud data is triangulated by the DelaunayTriangulation algorithm, and the generated triangular mesh Each pair of adjacent points in the point cloud is connected to form a preliminary model; Surface fitting and interpolation: Using known point cloud data, calculate and supplement missing surface points through interpolation methods , x, y, z are the three-dimensional coordinates of the missing surface point p, x, y represent the plane coordinates of the surface point on the surface of the plot, and z represents the ground height of the surface point, so that the surface morphology of the entire plot is smooth and continuous.

4. The method for automatic height adjustment of an intelligent weeder based on terrain height perception according to claim 3, characterized in that: The slope change trend of the area is calculated on the reconstructed three-dimensional model of the work plot, and the slope size is determined by calculating the direction of the ground normal vector. , the slope S is calculated as follows: ； Among them, arctan is the inverse tangent function, and They are height functions The partial derivatives in the x and y directions are discretized using the gradient operator. By traversing the grid of the entire operation plot, the slope value of each grid cell is obtained and the slope change trend is calculated. Extracted through statistical regression analysis.

5. The method for automatic height adjustment of an intelligent weeder based on terrain height perception according to claim 4, characterized in that: The gully distribution is displayed by marking all gully areas in the three-dimensional model of the working plot; Based on the ground points in the three-dimensional model of the work area, the gully depth of each gully area is calculated, and the laser point cloud data middle, is the ground height at the corresponding position, and the relative depth of each gully area is calculated using the ground height difference: , where z1 is the average height of the adjacent non-gully area and z2 is the height of the lowest point of the gully.

6. The method for automatic height adjustment of an intelligent lawn mower based on terrain height perception according to claim 5, characterized in that: The operation path planning in S3 adopts The algorithm prioritizes the gully area in combination with the terrain characteristics and sets the gully area as a high weight to ensure that the path planning prioritizes the gully area as part of the operation path. When calculating the path, for each operation point, the cost function to the target point is calculated to adjust the impact of various factors on the path selection.

7. The method for automatic height adjustment of an intelligent weeder based on terrain height perception according to claim 6, characterized in that: Said The algorithm specifically includes: S31, cost function definition: ensure that the weeding actuator enters the gully area first, design the cost function , the cost function reflects each grid point The difficulty level is: Slope cost: Areas with large slopes increase the difficulty of operations, so the slope cost is defined as: ,in is the slope weight coefficient, is the grid point The slope of the Gully cost: Gully areas have complex terrain and higher cost, defined as: ,in, is the weight coefficient of gully depth, is the grid point The depth of the gully at Total cost function Take the weighted sum of the two: ; S32, use The algorithm performs path search. The heuristic function of the algorithm Estimate the shortest distance from the current point to the target point, the cost function As the cost of the path, the goal is to find the path with the minimum total cost: ,in, is the actual cost from the starting point to the current grid point, the actual cost is the sum of the costs of all grid points from the starting point to the current grid point. The cost of each step comes from : , which means that each time you move from the starting point to a certain grid point , accumulate terrain cost along the path , For the heuristic estimate of the cost from the current grid point to the target grid point, choose the total cost The lowest path is taken as the optimal operation path.

8. The method for automatic height adjustment of an intelligent lawn mower based on terrain height perception according to claim 1, characterized in that: The target operating height of the weeding actuator in S4 Dynamically adjust according to the regional terrain characteristics of the current operation point: Flat area: If the current working grid point If the area is flat, the target working height is equal to the ground height. equal; Slope area: For areas with large slopes, Dynamically adjust the target working height to avoid collision with the ground or leaving the effective working range: ,in, is the slope adjustment factor; Gully Areas: For gully areas, the target working height should be relative to the lowest point of the gully Adjustments are made to ensure that the weeding actuator is adapted to the needs of deep trench operations: ,in, is the lowest point height of the current gully area, It is a preset safe working gap used to ensure the working distance of the weeding actuator.

9. The method for automatic height adjustment of an intelligent weeder based on terrain height perception according to claim 1, characterized in that: The S5 specifically includes: Based on the data provided by the ultrasonic sensor, the surface at the bottom of the gully is identified, and the inclination angle of the gully area is determined according to the gully depth information. The normal direction of the local surface at the bottom of the gully is set as , the direction is calculated as: , where z is at point The ground height of the location, and are the gradients of the point in the x and y directions respectively; By calculating the normal vector at the bottom of the gully And the original installation angle of the tool The angle between the two is used to determine the adjustment angle of the tool. Then, according to the direction of the normal vector, the target angle of the tool is calculated. : ,in, is the current direction vector of the tool, represents the dot product of vectors, and are the magnitudes of the vectors respectively.

Images