Intelligent navigation path planning method for cross-field operation of agricultural machinery in hilly and mountainous regions

Abstract

The invention discloses an intelligent navigation path planning method for cross-field operation of agricultural machinery in hills and mountains. The method comprises the steps that an unmanned aerial vehicle carrying an RGB camera carries out oblique photography operation on a terrace area at a hilly and mountain land to obtain oblique photography photos, POS data are obtained at the same time, an oblique photography model is constructed, coordinate conversion is carried out, then a digital surface model is established, a farmland area is cut off after preprocessing, and a terrace area map only containing a terrace path is obtained; and constructing an improved A * algorithm, and processing the terrace area map only containing the terrace path by using the improved A * algorithm to obtain an optimal driving path of the terrace path so as to realize intelligent navigation driving of the agricultural machinery in the hilly and mountainous regions. According to the method, the optimal driving path with the passing cost weight can be given according to the gradient, the slope angle and other information of the terrace area path, the working efficiency of the agricultural machinery in the hilly and mountainous regions is improved to a great extent, and the driving safety of the agricultural machinery is also ensured.

Description

Technical Field

[0001] The invention relates to a path planning method for hilly and mountainous areas, and in particular to an intelligent navigation path planning method for cross-field operations of agricultural machinery in hilly and mountainous areas. Background Art

[0002] The terrain of mountains and hills is complex and changeable, and traditional agricultural machinery often finds it difficult to adapt. However, unmanned vehicles can flexibly adapt to the changes in the slope and width of terraces through autonomous navigation and intelligent control systems, and perform tasks such as sowing, fertilizing, and weeding in terraces efficiently and continuously. Compared with traditional manual or mechanized operations, unmanned vehicles can significantly improve operation efficiency and shorten operation cycles. Unmanned vehicle operations reduce dependence on manpower, reduce labor intensity and agricultural production costs. The conversion between terrace plots involves complex operations such as uphill and downhill, and crossing gullies. During driving, unmanned vehicles need to perceive the surrounding environment in real time and accurately judge the position and speed of obstacles. However, features such as ridges and gullies of terraces may cause interference to GPS signals and affect positioning accuracy. In addition, visual sensors may also make misjudgments in environments with large light changes and many obstructions. At the same time, the paths between terrace plots are usually narrow and irregular, and unmanned vehicles need efficient path planning and navigation algorithms to plan safe driving routes. However, existing algorithms have difficulty in dealing with the particularity and complexity of terrace terrain, resulting in the planned paths being not optimized enough or unable to avoid obstacles. Summary of the invention

[0003] In order to solve the problems existing in the background technology, the present invention provides an intelligent navigation path planning method for cross-field operations of agricultural machinery in hilly and mountainous areas.

[0004] The technical solution adopted by the present invention is:

[0005] The method for intelligent navigation path planning for cross-field operation of agricultural machinery in hilly and mountainous areas of the present invention comprises:

[0006] S1: Oblique photography of the terraced field area in the hilly mountainous area is performed by using a drone equipped with an RGB camera to obtain several oblique photography photos of the terraced field area, and the position and orientation system POS (Position and Orientation System) data is obtained by using the drone at the same time.

[0007] S2: construct an oblique photography model of the terrace area according to each oblique photography photograph and the position and attitude system POS data; and establish a digital surface model DSM (Digital Surface Model) of the terrace area after coordinate conversion of the oblique photography model of the terrace area.

[0008] S3: After preprocessing the digital surface model DSM of the terrace area, the farmland area is cut out to obtain a terrace area map containing only the terrace paths.

[0009] S4: Construct an improved A* algorithm, and use the improved A* algorithm to process the terrace area map containing only terrace paths to obtain the optimal driving path of the terrace paths. When the hilly and mountainous agricultural machinery performs cross-field operations, it performs intelligent navigation according to the optimal driving path. The optimal planned path is exported to the GNSS map, and finally the driving tasks are preset and issued to the hilly and mountainous agricultural machinery.

[0010] In the step S2, an oblique photography model of the terraced field area is constructed by performing aerial triangulation and three-dimensional modeling in sequence according to each oblique photography photograph and the position and attitude system POS data.

[0011] In the step S2, the point cloud information in the oblique photography model of the terraced area is converted into a digital surface model DSM of the terraced area through coordinate transformation. The digital surface model DSM contains information such as the height and width of the terraced area, providing basic data for subsequent path planning.

[0012] In step S3, the digital surface model DSM of the terraced field area is pre-processed by performing denoising, filtering and coordinate correction in sequence.

[0013] In step S4, the improved A* algorithm is constructed by first establishing a travel cost weight table of slope angle and slope, as follows:

[0014]

[0015] Then the terrace area map containing only the terrace paths is segmented to obtain the slope k and slope angle α, and then the pass cost weight p is obtained according to the pass cost weight table. After the pass cost weight p is assigned to the A* algorithm, an improved A* algorithm is constructed.

[0016] When the terrace area map containing only terrace paths is segmented, the terrace area map containing only terrace paths is first segmented into a series of regular square grids as planning units. For each planning unit, the coordinate information of the four boundary points of the planning unit is obtained, thereby obtaining the elevations of the four boundary points, and obtaining the respective slopes k according to the elevations of the four boundary points, and selecting the largest slope k as the slope k of the planning unit. max , and then the slope k of the planning unit max Transformed into the slope angle α of the planning unit max .

[0017] The slope k of each boundary point is as follows:

[0018]

[0019] Among them, k i is the slope of the i-th boundary point; Z1, Z2, Z3 and Z4 are the elevations of the four boundary points respectively. The slope angle α of the planning unit max The details are as follows:

[0020] α max =arctan k max

[0021] After the travel cost weight p is assigned to the A* algorithm, the movement cost G in the A* algorithm is modified to obtain the modified movement cost G′, G′=[1+lg(1+p)]×G, thereby constructing an improved A* algorithm.

[0022] In the step S4, the terrace area map containing only terrace paths is processed using the improved A* algorithm to obtain the optimal driving path of the terrace paths. Specifically, accessibility is assigned to each planning unit using the travel cost weight p, and the planning units with accessibility values ​​D>1.75 are input as obstacles into the improved A* algorithm for path cost calculation, D=lg(1+p), so as to obtain the optimal planned path according to the improved A* algorithm.

[0023] The method of the present invention is particularly suitable for path selection of terraced field roads with complex terrain. First, through oblique photography, three-dimensional modeling and planning unit division slope assessment, the width, slope and gully information of the terraced road are accurately obtained. For the complex terrain of mountainous and hilly areas, the movement cost G in the A* algorithm is corrected, and the travel cost weight p is introduced to form a new path cost F'. The grid with particularly steep slopes is treated as an obstacle to enhance the adaptability under extreme terrain. The modified A* algorithm is used to calculate the optimal driving path from the starting point to the end point, and the optimal path is exported to the GNSS map. The driving tasks are preset for hilly and mountainous agricultural machinery to achieve intelligent and efficient path planning and execution.

[0024] The beneficial effects of the present invention are:

[0025] Compared with the traditional path planning method based on visual navigation, this method can effectively identify the elevation of field roads and calculate information such as slope and slope angle, thereby assigning the optimal path with a travel cost weight. In complex terrains such as mountains and hills, three-dimensional model navigation can provide accurate elevation information, helping users to easily find the target location and avoid potential danger areas, and to a certain extent can solve the elevation information ignored in traditional methods. The present invention effectively provides the travel cost weights between each grid and supplements the obstacle information based on this, and uses the modified A* algorithm to plan the optimal driving route for unmanned vehicles, which greatly improves the operating efficiency of agricultural machinery in hilly and mountainous areas and ensures its driving safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 is a flow chart of the method of the present invention;

[0027] Figure 2 This is a schematic diagram of the original A* algorithm;

[0028] Figure 3 The flow chart of the traffic weight and obstacle judgment of the present invention;

[0029] Figure 4 It is a schematic diagram of the improved A* algorithm of the present invention;

[0030] Figure 5 It is a comparison diagram of the best paths before and after correction of the present invention. DETAILED DESCRIPTION

[0031] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0032] like Figure 1 As shown, the intelligent navigation path planning method for cross-field operation of agricultural machinery in hilly and mountainous areas of the present invention is specifically as follows:

[0033] S1: Oblique photography of the terraced field area in the hilly mountainous area is performed by using a drone equipped with an RGB camera to obtain several oblique photography photos of the terraced field area, and the position and attitude system POS data is obtained by using the drone at the same time.

[0034] S2: Construct an oblique photography model of the terrace area based on each oblique photography photo and the position and attitude system POS data; specifically, import each oblique photography photo and the position and attitude system POS data into the contextcapture software, preset parameters, and then perform aerial triangulation and 3D modeling in sequence. The point cloud information in the oblique photography model of the terrace area is converted into a digital surface model DSM of the terrace area through coordinate conversion. The digital surface model DSM contains information such as the height and width of the terrace area, providing basic data for subsequent path planning.

[0035] S3: The digital surface model DSM of the terrace area is preprocessed by denoising, filtering and coordinate correction in sequence, and then the farmland area is cropped to obtain a terrace area map containing only the terrace paths.

[0036] S4: Construct an improved A* algorithm. First, establish a travel cost weight table for slope angle and slope, as follows:

[0037]

[0038] Then the terrace area map containing only the terrace paths is segmented to obtain the slope k and slope angle α, and then the pass cost weight p is obtained according to the pass cost weight table. After the pass cost weight p is assigned to the A* algorithm, an improved A* algorithm is constructed.

[0039] The terrace area map containing only terrace paths is processed using the improved A* algorithm to obtain the optimal driving path of the terrace paths. When the hilly and mountainous agricultural machinery is operating across fields, intelligent navigation is performed according to the optimal driving path. The optimal planned path is exported to the GNSS map, and finally the driving tasks are preset and issued to the hilly and mountainous agricultural machinery.

[0040] like Figure 3 As shown in FIG. 1 , when segmenting a terrace area map containing only terrace paths, the terrace area map containing only terrace paths is first segmented into a series of regular square grids with a length of d=0.2 m as planning units. For each planning unit, the spatial coordinates O(X, Y, Z5) of the center point of the planning unit and the coordinate information of the four boundary points are obtained to obtain the elevations of the four boundary points. The respective slopes k are obtained according to the elevations of the four boundary points, and the largest slope k is selected as the slope k of the planning unit. max , and then the slope k of the planning unit max Transformed into the slope angle α of the planning unit max .

[0041] The slope k of each boundary point is as follows:

[0042]

[0043] Among them, k i is the slope of the i-th boundary point; Z1, Z2, Z3 and Z4 are the elevations of the four boundary points respectively. The slope angle α of the planning unit max The details are as follows:

[0044] α max =arctan k max

[0045] After the travel cost weight p is assigned to the A* algorithm, the movement cost G in the A* algorithm is modified to obtain the modified movement cost G′, G′=[1+lg(1+p)]×G, thereby constructing an improved A* algorithm.

[0046] like Figure 4 As shown, the improved A* algorithm can obtain a new path cost F′ according to the modified movement cost G′, as follows:

[0047] F′=G′+H=[1+lg(1+p)]×G+H

[0048] Figure 4In the figure, the numbers in the lower left and lower right corners of each grid are the movement cost G in the A* algorithm and the estimated cost H of moving from the specified square to the end point B grid.

[0049] like Figure 2 As shown, in the original A* algorithm, the path cost F between the starting point A grid and the end point B grid can be obtained by the following formula:

[0050] F=G+H

[0051] Among them, F is the path cost obtained by the A* algorithm, G is the movement cost of moving from the starting point A grid to the specified grid, and the path generated along the way to reach the grid; H is the estimated cost of moving from the specified grid to the end point B grid.

[0052] Figure 2 In the figure, the numbers in the lower left and right corners of each grid are the modified moving cost G′ in the improved A* algorithm and the estimated cost H of moving from the specified square to the end point B grid.

[0053] The terrace area map containing only terrace paths is processed by the improved A* algorithm to obtain the optimal driving path of the terrace paths. Specifically, accessibility is assigned to each planning unit using the travel cost weight p, and the planning units with accessibility values ​​D>1.75 are input as obstacles into the improved A* algorithm for path cost calculation, D=lg(1+p), so as to obtain the optimal planned path according to the improved A* algorithm.

[0054] like Figure 5 As shown, it can be seen that the method of the present invention plans an optimal driving route with a shorter path under the premise of being able to avoid obstacles, which greatly improves the operating efficiency of agricultural machinery in hilly and mountainous areas and ensures its driving safety.

Claims

1. An intelligent navigation path planning method for cross-field operation of agricultural machinery in hilly and mountainous areas, characterized in that: include: S1: Oblique photography of terraced fields in hilly areas is performed by drones equipped with RGB cameras to obtain several oblique photographs of the terraced fields, and position and attitude system POS data is obtained by drones at the same time; S2: construct an oblique photography model of the terrace area based on each oblique photography photo and the position and attitude system POS data; and establish a digital surface model DSM of the terrace area after coordinate conversion of the oblique photography model of the terrace area; S3: After preprocessing the digital surface model DSM of the terrace area, the farmland area is cut out to obtain a terrace area map containing only the terrace paths; S4: Construct an improved A* algorithm, and use the improved A* algorithm to process the terrace area map that only contains terrace paths to obtain the optimal driving path of the terrace paths. When agricultural machinery in hilly and mountainous areas performs cross-field operations, intelligent navigation is performed according to the optimal driving path.

2. The intelligent navigation path planning method for cross-field operation of agricultural machinery in hilly and mountainous areas according to claim 1 is characterized by: In the step S2, an oblique photography model of the terraced field area is constructed by performing aerial triangulation and three-dimensional modeling in sequence according to each oblique photography photograph and the position and attitude system POS data.

3. The intelligent navigation path planning method for cross-field operation of agricultural machinery in hilly and mountainous areas according to claim 1 is characterized by: In the step S2, the point cloud information in the oblique photography model of the terraced field area is converted into a digital surface model DSM of the terraced field area through coordinate transformation, and the digital surface model DSM includes the height and width information of the terraced field area.

4. The intelligent navigation path planning method for cross-field operation of agricultural machinery in hilly and mountainous areas according to claim 1 is characterized by: In step S3, the digital surface model DSM of the terraced field area is pre-processed by performing denoising, filtering and coordinate correction in sequence.

5. The intelligent navigation path planning method for cross-field operation of agricultural machinery in hilly and mountainous areas according to claim 1 is characterized by: In step S4, the improved A* algorithm is constructed by first establishing a travel cost weight table of slope angle and slope, as follows: Then the terrace area map containing only the terrace paths is segmented to obtain the slope k and slope angle α, and then the pass cost weight p is obtained according to the pass cost weight table. After the pass cost weight p is assigned to the A* algorithm, an improved A* algorithm is constructed.

6. The intelligent navigation path planning method for cross-field operation of agricultural machinery in hilly and mountainous areas according to claim 5 is characterized by: When the terrace area map containing only terrace paths is segmented, the terrace area map containing only terrace paths is first segmented into a series of regular square grids as planning units. For each planning unit, the coordinate information of the four boundary points of the planning unit is obtained, thereby obtaining the elevations of the four boundary points, and the respective slopes k are obtained according to the elevations of the four boundary points, and the largest slope k is selected as the slope k of the planning unit. max , and then the slope k of the planning unit max Transformed into the slope angle α of the planning unit max .

7. The intelligent navigation path planning method for cross-field operation of agricultural machinery in hilly and mountainous areas according to claim 6 is characterized by: After the travel cost weight p is assigned to the A* algorithm, the movement cost G in the A* algorithm is modified to obtain the modified movement cost G′, G′=[1+lg(1+p)]×G, thereby constructing an improved A* algorithm.

8. The intelligent navigation path planning method for cross-field operation of agricultural machinery in hilly and mountainous areas according to claim 6 is characterized by: In the step S4, the terrace area map containing only terrace paths is processed using the improved A* algorithm to obtain the optimal driving path of the terrace paths. Specifically, accessibility is assigned to each planning unit using the travel cost weight p, and the planning units with accessibility values ​​D>1.75 are input as obstacles into the improved A* algorithm, D=lg(1+p), so as to obtain the optimal planned path according to the improved A* algorithm.

Images