A method for generating variable fertilization paths for unmanned plant protection vehicles in orchards based on remote sensing technology

Abstract

The present invention discloses a method for generating a variable fertilization path for an unmanned plant protection vehicle in an orchard based on remote sensing technology, comprising the following steps: S1, data acquisition and processing; S2, calculation of the NDVI index and extraction of the fruit tree canopy; S3, extraction of the centroid; S4, centroid fitting; S5, calculation of the trajectory line between fruit tree rows; S6, generation of a U-turn trajectory; S7, calculation of the vegetation index density ViDPT index of each tree; S8, calculation of a fertilization prescription map for each tree; S9, generation of a driving path with fertilization information; S10, uploading the driving path with fertilization information to the plant protection vehicle to achieve variable fertilization. The present invention can quickly complete the setup work for the plant protection vehicle. Using a drone, only one shot of the orchard is needed to quickly complete the operation path planning of the orchard plant protection vehicle, eliminating the need for tedious process settings and saving manpower and time. Based on the path generation method of the present invention, the plant protection vehicle can perform variable fertilization according to the growth conditions of the plants, thereby saving fertilizer and reducing planting costs.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent agriculture and relates to a method for generating a variable fertilization path for an unmanned plant protection vehicle in an orchard based on remote sensing technology. Background Art

[0002] At present, conventional plant protection operations generally use a single continuous spraying operation with the same fertilizer amount in the operation area. Since this operation method does not take into account the growth conditions and differences of crops in the operation area, that is, it cannot be adjusted according to the growth conditions of plants, it is easy to cause insufficient fertilizer and pesticide use in areas with good growth conditions, and excessive use in areas with poor growth conditions, resulting in fertilizer and pesticide waste and increased planting costs.

[0003] The traditional path generation method requires the user to manually control the plant protection vehicle to record the path on site in the operation area. The path is generated using the AB line method. The plant protection vehicle is parked at the starting point of a row of fruit trees, with the front of the vehicle facing the same direction as the row of fruit trees. The combined navigation provided by the plant protection vehicle is used to record the current point coordinates as point A. The plant protection vehicle is then parked at the end of the row of fruit trees, and the current point coordinates are recorded as point B to generate a planned path. This method has a complex planning process, and the entire planning process is manually performed. Personnel are required to manually drive and record the driving trajectory. There cannot be interruptions or operational errors in the middle, and in actual operations, it is often necessary to record the plan repeatedly. However, this method is only applicable to situations where the planting arrangement of fruit trees is horizontal and vertical. If the fruit trees are planted in mountainous and hilly areas, and the fruit trees are not planted in a straight line, the AB line generation method cannot be applied. Summary of the Invention

[0004] In view of the technical problems existing in the existing methods, the purpose of the present invention is to provide a method for generating a variable fertilization path for an unmanned plant protection vehicle in an orchard based on remote sensing technology.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] A method for generating a variable fertilization path for an unmanned plant protection vehicle in an orchard based on remote sensing technology includes the following steps:

[0007] S1. Data acquisition and processing: Survey and photograph the operation area to obtain orchard image information, and then perform ortho-mosaic processing on the obtained image information to obtain the planning base map;

[0008] S2. Calculate the NDVI index (i.e., Normalized Difference Vegetation Index) and extract the canopy of fruit trees:

[0009] Using the near-infrared band image and the red band image obtained after the orthomosaic processing in step S1, the NDVI index is calculated according to formula (1), thereby obtaining the NDVI base map of the operation area;

[0010]

[0011] Among them, NIR represents near-infrared band images; Red represents red light band images;

[0012] The OpenCV image processing library is then used to identify the outlines of fruit trees on the obtained NDVI base map and extract the fruit tree canopy as a mask. The specific steps include: extracting the target color, analyzing the color characteristics of the fruit trees in the orchard, determining the color range of the fruit trees, distinguishing the fruit trees from the background, and obtaining the fruit tree canopy as a mask;

[0013] S3. Extract centroids: Use the "findContours" function in OpenCV to find the centroids, record the positions of the centroids, and obtain the centroid queue for each row of trees.

[0014] S4. Centroid fitting: Use the least squares method to fit the centroid queue of each row of trees to obtain the fitting line where the fruit tree row is located;

[0015] S5. Calculate the trajectory line between fruit tree rows: Take the parallel midlines of the fitted lines of two adjacent fruit tree rows to obtain the trajectory line of the plant protection vehicle traveling in the middle of the fruit tree row;

[0016] S6. Generate a U-turn trajectory: Use a Bezier curve to connect two adjacent trajectory lines in a "bow" shape end to end to generate a U-turn trajectory. The two endpoints of the curve are the endpoints of the two adjacent trajectory lines respectively. The curve radius r is determined based on the actual turning radius R of the plant protection vehicle and the distance D between the two adjacent trajectory lines. After connecting the endpoints of the two adjacent trajectory lines, the walking trajectory of the plant protection vehicle is obtained;

[0017] S7. Calculate the vegetation index density per tree (ViDPT index): Use the fruit tree canopy extracted in step S2 as a mask to extract the NDVI base map to obtain the NDVI image of the fruit tree canopy. Use the "boundingRect" function to calculate the circumscribed rectangle rot (rectangle of tree) of each fruit tree canopy. Then calculate the vegetation index density per tree (ViDPT index) according to formula (2):

[0018]

[0019] in, Indicates the accumulation of NDVI values ​​of each fruit tree canopy; S rot Represents the area of ​​the circumscribed rectangle rot of each fruit tree canopy;

[0020] S8. Calculate the fertilizer prescription map for each tree: Calculate the fertilizer amount Q for each tree according to formula (3) using all the obtained ViDPT indices:

[0021] Q=k*ViDPT*C (3)

[0022] Among them, Q represents the amount of fertilizer applied, k represents the deposition efficiency summarized based on experience, and C represents the minimum deposition threshold of each liquid fertilizer, which can ensure the coverage rate of fertilizer and pesticide.

[0023] Then, the fertilizer amount Q of each tree is overlaid on the circumscribed rectangular rot area of ​​each tree to obtain a fertilizer prescription map based on the ViDPT index;

[0024] S9. Sampling on the fertilizer prescription map according to the plant protection vehicle's driving path to generate a driving path with fertilization information: Use the "Generate Points Along Lines" function to convert the plant protection vehicle's driving trajectory obtained in step S6 into coordinate points, calculate the points R / 2 away from each coordinate point on the left and right sides, and sample according to the fertilizer prescription map to obtain the fertilizer amount on the left and right sides of the plant protection vehicle when it reaches the point, thereby generating a driving path with fertilization information;

[0025] S10. The driving route with fertilization information is uploaded to the plant protection vehicle. During actual operation, whenever the vehicle reaches a point, the fertilizer amount on both sides of the point is read to perform fertilization operation. The points are read continuously during the driving process to achieve variable-rate fertilization.

[0026] Preferably, in step S1, a drone equipped with a multispectral camera is used to survey and photograph the operating area to obtain orchard image information. In step S1, a drone equipped with a multispectral camera is used to survey and photograph the operating area to obtain orchard image information.

[0027] Preferably, in step S2, the OpenCV image processing library is used to identify the outline of the fruit tree and extract the fruit tree canopy. The specific steps include: extracting the target color, analyzing the color characteristics of the fruit trees in the orchard, determining the color range of the fruit trees, and distinguishing the fruit trees from the background; then performing opening and closing operations to remove noise and interference, and generating a mask, i.e., the fruit tree canopy.

[0028] Preferably, in step S6, the curve radius r is determined based on the actual turning radius R of the plant protection vehicle and the distance D between two adjacent trajectory lines. The specific principle is: if the trajectory line distance D is greater than twice the actual turning radius of the plant protection vehicle 2R, then r=D; if the distance D between two adjacent trajectory lines is less than 2R, then 2R is used as the radius r of the Bezier curve.

[0029] Beneficial effects:

[0030] Compared with the current plant protection vehicles, the method of the present invention can quickly complete the setting work for the plant protection vehicle. Using a drone, it only needs to take a photo of the orchard once to quickly complete the planning of the plant protection vehicle's operating path in the orchard, eliminating the need for cumbersome process settings and saving manpower and time. The present invention samples the plant protection vehicle's driving path on the vegetation index map to obtain a driving path with semantics (fertilization information), and can achieve the purpose of variable fertilization of the plant protection vehicle according to the growth conditions of the plants without modifying or adding sensors, saving fertilizer and reducing planting costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 This is a schematic diagram of the process framework of the variable fertilization path generation method for orchard unmanned plant protection vehicles based on remote sensing technology;

[0032] Figure 2 is the result of extracting the centroid;

[0033] Figure 3 This is the centroid point fitting result graph;

[0034] Figure 4 Schematic diagram of two U-turn methods;

[0035] Figure 5 This is a graph showing the ViDPT index results calculated in Example 1;

[0036] Figure 6 is the fertilization prescription map based on the ViDPT index calculated in Example 1;

[0037] Figure 7 Generate a driving route map with fertilization information in Example 1. DETAILED DESCRIPTION

[0038] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0039] Example 1

[0040] See Figure 1-7 The method for generating a variable fertilization path for an orchard unmanned plant protection vehicle based on remote sensing technology includes the following steps:

[0041] S1. Data acquisition and processing: Survey and photograph the operation area to obtain orchard image information, and then perform ortho-mosaic processing on the obtained image information to obtain the planning base map;

[0042] Specifically, drones equipped with multispectral cameras are used to survey and photograph the work area, acquiring image information of large orchards. This method of acquiring orchard images using drones is efficient and flexible. Satellite remote sensing imagery can also be used instead of drones to acquire orchard image information. The present invention uses ortho-mosaicing to process orchard image information, eliminating image distortion and tilt, enabling the planning base map to more accurately reflect the actual orchard terrain and fruit tree distribution.

[0043] S2. Calculation of the NDVI index (i.e., Normalized Difference Vegetation Index) and extraction of the fruit tree canopy: Using the near-infrared band image and the red light band image obtained after ortho-mosaic processing in step S1, the NDVI index is calculated according to formula (1), and the NDVI base map of the operation area is obtained:

[0044]

[0045] Among them, NIR represents near-infrared band images; Red represents red light band images;

[0046] The OpenCV image processing library is then used to identify the outlines of fruit trees on the obtained NDVI base map and extract the fruit tree canopy as a mask. The specific steps include: extracting the target color, analyzing the color characteristics of the fruit trees in the orchard, determining the color range of the fruit trees to distinguish the fruit trees from the background, and then performing opening and closing operations to remove noise and interference, making the fruit tree outlines clearer, and extracting the fruit tree canopy as a mask;

[0047] S3. Extract centroids: Use the "findContours" function in OpenCV to find the centroids, record their positions, and obtain a centroid queue for each row of trees. Here, the centroid queue represents the location information of the fruit tree rows.

[0048] S4. Centroid fitting: Use the least squares method to fit the centroid queue of each row of trees to obtain the fitting line where the fruit tree row is located;

[0049] S5. Calculate the trajectory line between fruit tree rows: Take the parallel midlines of the fitted lines of two adjacent fruit tree rows to obtain the trajectory line of the plant protection vehicle traveling in the middle of the fruit tree row;

[0050] S6. Generate a U-turn trajectory: Use a Bezier curve to connect two adjacent trajectory lines in a "bow" shape end to end to generate a U-turn trajectory. The two endpoints of the curve are the endpoints of the two adjacent trajectory lines. The curve radius r is determined based on the actual turning radius R of the plant protection vehicle and the distance D between the two adjacent trajectory lines: if the trajectory line distance D is greater than twice the actual turning radius of the plant protection vehicle 2R, then r = D; if the distance D between the two adjacent trajectory lines is less than 2R, then 2R is used as the radius r of the Bezier curve; then the endpoints of the two adjacent trajectory lines are connected to obtain the walking trajectory of the plant protection vehicle;

[0051] S7. Calculate the vegetation index density per tree (ViDPT index):

[0052] The fruit tree canopy extracted in step S2 is used as a mask to extract the NDVI base map to obtain the NDVI image of the fruit tree canopy. The “boundingRect” function is used to calculate the circumscribed rectangle rot (rectangle of tree) of each fruit tree canopy. Then, the vegetation index density ViDPT index of each tree is calculated according to formula (2):

[0053]

[0054] in, Indicates the accumulation of NDVI values ​​of each fruit tree canopy; S rot Represents the area of ​​the circumscribed rectangle rot of each fruit tree canopy;

[0055] S8. Calculate the fertilizer prescription map for each tree: Calculate the fertilizer amount Q for each tree according to formula (3) using all the obtained ViDPT indices:

[0056] Q=k*ViDPT*C (3)

[0057] Among them, Q represents the amount of fertilizer applied, k represents the deposition efficiency summarized based on experience, and C represents the minimum deposition threshold of each liquid fertilizer, which can ensure the coverage rate of fertilizer and pesticide.

[0058] Then, the fertilizer amount Q of each tree is overlaid on the circumscribed rectangular rot area of ​​each tree to obtain a fertilizer prescription map based on the ViDPT index;

[0059] S9. Sampling on the fertilizer prescription map according to the plant protection vehicle's driving path to generate a driving path with fertilization information: Use the "Generate Points Along Lines" function to convert the plant protection vehicle's driving trajectory obtained in step S6 into coordinate points, calculate the points R / 2 away from each coordinate point on the left and right sides, and sample them on the fertilization prescription map to obtain the fertilizer amount on the left and right sides of the plant protection vehicle when it reaches the point, thereby generating a driving path with fertilization information;

[0060] S10. The driving route with fertilization information is uploaded to the plant protection vehicle. During actual operation, whenever the vehicle reaches a point, the fertilizer amount on both sides of the point is read to perform fertilization operation. The points are read continuously during the driving process to achieve variable-rate fertilization.

[0061] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes based on the technical solutions and improved concepts of the present invention within the technical scope disclosed by the present invention, and these changes should be covered by the scope of protection of the present invention.

Claims

1. A method for generating variable-rate fertilization paths for orchard unmanned plant protection vehicles based on remote sensing technology, characterized by: The steps include: S1. Data acquisition and processing: Survey and photograph the operation area to obtain orchard image information, and then perform ortho-mosaic processing on the obtained image information to obtain the planning base map; S2. Calculation of NDVI index and extraction of fruit tree canopy: Using the near-infrared band image and the red light band image obtained after ortho-mosaic processing in step S1, calculate the NDVI index according to formula (1) to obtain the NDVI base map of the operation area: Among them, NIR represents near-infrared band images; Red represents red light band images; The OpenCV image processing library is then used to identify the outlines of fruit trees on the resulting NDVI base map and extract the tree canopy as a mask. This involves extracting the target color and analyzing the color characteristics of the fruit trees in the orchard to determine the color range of the fruit trees, thereby distinguishing the fruit trees from the background. S3. Extract centroids: Use the "findContours" function in OpenCV to find the centroids, record the positions of the centroids, and obtain the centroid queue for each row of trees. S4. Centroid fitting: Use the least squares method to fit the centroid queue of each row of trees to obtain the fitting line where the fruit tree row is located; S5. Calculate the trajectory line between fruit tree rows: Take the parallel midlines of the fitted lines of two adjacent fruit tree rows to obtain the trajectory line of the plant protection vehicle traveling in the middle of the fruit tree row; S6. Generate a U-turn trajectory: Use a Bezier curve to connect two adjacent trajectory lines in a "bow" shape. The two endpoints of the curve are the endpoints of the two adjacent trajectory lines. The curve radius r is determined based on the actual turning radius R of the plant protection vehicle and the distance D between the two adjacent trajectory lines. Then, the endpoints of the two adjacent trajectory lines are connected to obtain the travel trajectory of the plant protection vehicle. S7. Calculate the vegetation index density ViDPT index of each tree: Use the fruit tree canopy extracted in step S2 as a mask to extract the NDVI base map to obtain the NDVI image of the fruit tree canopy, and use the "boundingRect" function to calculate the circumscribed rectangle rot of each fruit tree canopy, and then calculate the vegetation index density ViDPT index of each tree according to formula (2): in, Indicates the accumulation of NDVI values ​​of each fruit tree canopy; S rot Represents the area of ​​the circumscribed rectangle rot of each fruit tree canopy; S8. Calculate the fertilizer prescription map for each tree: Calculate the fertilizer amount Q for each tree according to formula (3) using all the obtained ViDPT indices: Q=k*ViDPT*C (3) Among them, Q represents the amount of fertilizer applied, k represents the deposition efficiency based on experience, and C represents the minimum deposition threshold of each liquid fertilizer; Then, the fertilizer amount Q of each tree is overlaid on the circumscribed rectangular rot area of ​​each tree to obtain a fertilizer prescription map based on the ViDPT index; S9. Sample the fertilizer prescription map based on the plant protection vehicle's driving path to generate a driving path with fertilization information: Use the "Generate Points Along Lines" function to convert the plant protection vehicle's driving trajectory obtained in step S6 into coordinate points. Calculate the points R / 2 away from each coordinate point on the left and right sides. Sampling is performed based on the fertilizer prescription map to obtain the fertilizer amount on both sides of the plant protection vehicle when the vehicle reaches that point, thereby generating a driving path with fertilization information. S10, variable fertilization: The driving route with fertilization information is uploaded to the plant protection vehicle. During actual operation, whenever the vehicle reaches a point, the fertilizer amount on both sides of the point is read to perform fertilization operations. The points are read continuously during the driving process to achieve variable fertilization.

2. The method for generating a variable-rate fertilization path for an unmanned plant protection vehicle for an orchard based on remote sensing technology according to claim 1, characterized in that: In step S1, a drone equipped with a multispectral camera is used to survey and photograph the operation area to obtain orchard image information.

3. The method for generating a variable-rate fertilization path for an unmanned plant protection vehicle for an orchard based on remote sensing technology according to claim 2, characterized in that: In step S2, the OpenCV image processing library is used to identify the outline of the fruit tree and extract the fruit tree canopy. The specific steps include: extracting the target color, analyzing the color characteristics of the fruit trees in the orchard, determining the color range of the fruit trees, and distinguishing the fruit trees from the background; then performing opening and closing operations to remove noise and interference, and generating a mask, i.e., the fruit tree canopy.

4. The method for generating a variable-rate fertilization path for an unmanned plant protection vehicle for an orchard based on remote sensing technology according to claim 1, characterized in that: In step S6, the principle of determining the curve radius r based on the actual turning radius R of the plant protection vehicle and the distance D between two adjacent track lines is: if the track line distance D is greater than twice the actual turning radius of the plant protection vehicle 2R, then r=D; if the distance D between two adjacent track lines is less than 2R, then 2R is used as the radius r of the Bezier curve.

Images