A fritillaria thunbergii disease and pest control spraying path intelligent planning method

Abstract

This invention relates to the field of intelligent plant protection technology, specifically to an intelligent planning method for spraying paths to control pests and diseases of Fritillaria thunbergii (Zhejiang fritillary bulb). The method includes: acquiring thermal imaging images of the Zhejiang fritillary bulb planting area and the spraying planning path of a drone; identifying conflict areas based on the distribution of the number of drone spraying planning paths at each grid area location and the pixel values ​​in the thermal imaging images; obtaining a flight difficulty index based on the spatiotemporal distance between adjacent grid areas on each drone's spraying planning path, combined with the pixel values ​​of each conflict area in the thermal imaging images; and replanning the drone spraying planning path based on the changes in the flight difficulty index of each conflict area and the flight difficulty index of other grid areas on each drone's spraying planning path. This invention avoids spatial conflicts during drone flight caused by local path optimization, improving the rationality of the drone path planning scheme.

Description

Technical Field

[0001] This invention relates to the field of intelligent plant protection technology, specifically to an intelligent planning method for spraying paths to control pests and diseases of Fritillaria thunbergii. Background Technology

[0002] With the widespread adoption of precision agriculture technology, agricultural drones have become key equipment for the green control of pests and diseases in large-scale farmland due to their core advantages such as high operational efficiency, strong terrain adaptability, and high precision in pesticide application. Their application is particularly widespread in the large-scale cultivation of specialty economic crops such as Fritillaria thunbergii. However, the unique characteristics of Fritillaria thunbergii cultivation areas (mostly distributed on slopes and hills with complex terrain; short plants (30-50cm), dense planting, and sensitivity to pesticides), along with the practical need for multi-drone collaborative operations, have created bottlenecks for traditional drone path planning algorithms.

[0003] Traditional path planning algorithms, such as ant colony optimization, aim to minimize the flight distance of a single drone, guiding multiple drones to prioritize flat terrain and areas close to the base station (such as flat ground at the bottom of a slope). This leads to these areas becoming high-frequency conflict zones. However, the risk of pests and diseases in Fritillaria thunbergii may be concentrated in the middle of the slope (where high humidity increases susceptibility to disease), and traditional algorithms cannot balance distance with pest and disease priority. Therefore, existing spraying path planning methods are prone to overlapping paths between multiple drones, resulting in poor path planning rationality. Summary of the Invention

[0004] To address the technical problem of poor path planning rationality caused by overlapping paths of multiple sprayers in existing spraying path planning methods, the present invention aims to provide an intelligent path planning method for spraying pests and diseases of Fritillaria thunbergii. The specific technical solution adopted is as follows:

[0005] Acquire thermal imaging images of the Fritillaria thunbergii planting area and the spraying planning path of each UAV, wherein the spraying planning path includes several grid areas of the Fritillaria thunbergii planting area;

[0006] Based on the distribution of the number of grid regions belonging to the drone spraying planning path at the location of each grid region, and the pixel value of each grid region in the thermal imaging image, the grid regions are filtered to obtain conflict areas.

[0007] Based on the spatiotemporal distance distribution between each grid area and adjacent grid areas on the spraying planning path of each UAV, and combined with the pixel value of each conflict area in the thermal imaging image, the flight difficulty index of each grid area on the spraying planning path of each UAV is obtained.

[0008] Based on the changes in the flight difficulty index of each conflict area and the flight difficulty index of other grid areas along the spraying planning path of each drone, the drone spraying planning path is replanned.

[0009] Preferably, the step of filtering grid regions to obtain conflict areas based on the distribution of the number of grid regions belonging to the UAV spraying planning path at the location of each grid region and the pixel value of each grid region in the thermal imaging image specifically includes:

[0010] The proportion of the number of spraying planning paths of different drones at the location of each grid area is obtained, and the ratio of the proportion of the number to the pixel value of each grid area in the thermal imaging area is used as the conflict feature index of each grid area.

[0011] Based on the conflict characteristic index of each grid region, the grid regions are filtered to obtain conflict areas.

[0012] Preferably, the step of filtering the grid regions based on the conflict characteristic index of each grid region to obtain conflict regions specifically includes:

[0013] The grid area corresponding to the conflict feature index being greater than or equal to the preset conflict threshold is defined as the conflict area.

[0014] Preferably, the step of obtaining the flight difficulty index of each grid area on the spraying planning path of each drone based on the spatiotemporal distance distribution between each grid area and adjacent grid areas on the spraying planning path of each drone, combined with the pixel value of each conflict area in the thermal imaging image, specifically includes:

[0015] Based on the flight distance distribution between the selected grid area and adjacent grid areas on each drone spraying planning path, the flight cost factor of the selected grid area on each drone spraying planning path is obtained.

[0016] The sum of the differences between the flight cost factor of the selected grid area on the selected spraying planning path and the flight cost factor of the other UAV spraying planning paths is used as the cost fluctuation degree of the selected grid area. The ratio of the flight cost factor of the selected grid area on the selected spraying planning path to the cost fluctuation degree is used as the spatiotemporal reconstruction parameter of the selected grid area on the selected spraying planning path.

[0017] The selected grid area is any grid area on the spray planning path; the selected spray planning path is any spray planning path of any UAV.

[0018] Based on the spatiotemporal distance distribution between the selected grid area and adjacent grid areas on the spraying planning path, the pixel values ​​in the thermal imaging image, and the spatiotemporal reconstruction parameters, the flight difficulty index of the selected grid area is obtained.

[0019] Preferably, the step of obtaining the flight cost factor of the selected grid area on each drone spraying planning path based on the flight distance distribution between the selected grid area and adjacent grid areas on each drone spraying planning path specifically includes:

[0020] Obtain the drone flight distance for each grid area along the spraying plan path;

[0021] For any drone's spraying planning path, the absolute value of the difference between the drone's flight distance between the selected grid area and the first adjacent grid area is obtained as the first flight difference; the absolute value of the difference between the drone's flight distance between the selected grid area and the second adjacent grid area is obtained as the second flight difference; the sum of the first flight difference and the second flight difference is the flight cost factor of the selected grid area.

[0022] The first adjacent grid area is the preceding grid area adjacent to the selected grid area on the spraying planning path, and the second adjacent grid area is the following grid area adjacent to the selected grid area on the spraying planning path.

[0023] Preferably, obtaining the flight difficulty index of the selected grid area based on the spatiotemporal distance distribution between the selected grid area and adjacent grid areas on the spraying planning path, the pixel values ​​in the thermal imaging image, and the spatiotemporal reconstruction parameters specifically includes:

[0024] Based on the time corresponding to the selected grid area on the spraying planning path and the location of the selected grid area, the spatiotemporal distance between the selected grid area and the second adjacent grid area is determined.

[0025] The ratio between the pixel value of the selected grid region in the thermal imaging image and the product of the spatiotemporal distance and the spatiotemporal reconstruction parameter is used as the flight difficulty index of the selected grid region.

[0026] Preferably, the step of replanning the drone spraying path based on the changes in the flight difficulty index of each conflict area and the flight difficulty index of other grid areas along the spraying planning path of each drone specifically includes:

[0027] For any given spraying plan path, obtain the non-conflict areas other than the target conflict area as candidate grid areas; the target conflict area is any conflict area on the spraying plan path.

[0028] Based on the difference in flight difficulty index between the target conflict area and each candidate grid area, the path reconstruction parameters for each candidate grid area are obtained;

[0029] Based on the path reconstruction parameters of each candidate grid area, the spraying planning path is adjusted to obtain the reconstructed path of the UAV spraying planning path.

[0030] Preferably, obtaining the path reconstruction parameters for each candidate grid region based on the difference in flight difficulty index between the target conflict region and each candidate grid region specifically includes:

[0031] The difference in flight difficulty index between the target conflict area and each candidate grid area is used as the path reconstruction parameter for each candidate grid area.

[0032] Preferably, the step of adjusting the spraying planning path based on the path reconstruction parameters of each candidate grid region to obtain the reconstructed path of the UAV spraying planning path specifically includes:

[0033] The candidate raster region corresponding to the path reconstruction parameter being greater than the preset reconstruction threshold and the maximum value of the path reconstruction parameter is selected as the adjustment raster region.

[0034] The target conflict area in the spraying planning path is adjusted to an adjustment grid area. The grid area after the target conflict area in the adjusted spraying planning path is then replanned to obtain the reconstructed path.

[0035] Preferably, the method for obtaining the spraying planning path for each UAV specifically includes:

[0036] The drone flight distance for each grid area is obtained, and the average pixel value of all pixels in the thermal image of each grid area is used as the pest and disease intensity of each grid area.

[0037] The negative correlation coefficient of the drone's flight distance is used as the pheromone for the ant colony algorithm, and the intensity of the pests and diseases is used as the priority weight. The ant colony algorithm is used to plan the flight path of the drones to obtain the spraying plan path for each drone.

[0038] The embodiments of the present invention have at least the following beneficial effects:

[0039] This invention first distinguishes between reasonable repetition and invalid conflict, quantifies the degree of unreasonableness of the conflict, and provides a precise adjustment target for subsequent path reconstruction and reduction of repeated spraying costs.

[0040] Then, by analyzing the spatiotemporal distance between each grid area and adjacent grid areas along the drone's spraying planning path, and combining this with pixel distribution in thermal imaging images, the severity of pests and diseases is assessed, thereby quantifying the overall cost pressure of the drone flying in a specific grid area. Finally, by analyzing the changes in flight difficulty in other grid areas and conflict areas, the drone flight path is optimized. This invention separates path conflicts generated by multiple drones during flight, ensuring mutual coordination among multiple drones during flight. Through individual drone path optimization and collaborative verification, global conflicts are reduced, perfectly adapting to the needs of multi-drone collaborative spraying in Fritillaria thunbergii cultivation areas. This invention, to a certain extent, avoids spatial conflicts during drone flight caused by local path optimization, improving the rationality of the drone path planning scheme. Attached Figure Description

[0041] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0042] Figure 1 This is a flowchart of the steps of an intelligent planning method for spraying pesticide paths to control pests and diseases of Fritillaria thunbergii provided by the present invention;

[0043] Figure 2 This is a flowchart of the steps of the method for obtaining conflict regions provided by the present invention;

[0044] Figure 3 This is a flowchart of the steps for obtaining flight difficulty indicators provided by the present invention;

[0045] Figure 4 This is a flowchart of the sub-steps of step S400 provided by the present invention. Detailed Implementation

[0046] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a method for intelligent planning of spraying paths for the prevention and control of diseases and pests of Fritillaria thunbergii according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0047] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0048] The following description, in conjunction with the accompanying drawings, details the specific scheme of the intelligent planning method for spraying paths for the prevention and control of diseases and pests of Fritillaria thunbergii provided by this invention.

[0049] Please see Figure 1 The diagram illustrates a flowchart of a method for intelligent planning of spraying paths for pest and disease control of Fritillaria thunbergii, according to an embodiment of the present invention. The method includes the following steps:

[0050] Step S100: Obtain thermal imaging images of the Fritillaria thunbergii planting area and the spraying planning path of each UAV, wherein the spraying planning path includes several grid areas of the Fritillaria thunbergii planting area.

[0051] First, remote sensing data of the Fritillaria thunbergii planting area was acquired via high-resolution satellites, including digital elevation data (DEM), digital surface data (DSM), and vegetation cover images. The vegetation cover images generally include RGB, multispectral, and hyperspectral images. It should be understood that DEM reflects the absolute height of the terrain itself (excluding vegetation), while DSM reflects the absolute height of the land surface (including the vegetation canopy). The difference between the DEM and DSM represents the vegetation height within the Fritillaria thunbergii planting area. It should be noted that the methods for acquiring DEM and DSM are well-known technologies and will not be elaborated upon further here.

[0052] Then, using drones equipped with infrared thermal imaging devices, thermal images of the Fritillaria thunbergii planting area were acquired to analyze the extent of pest and disease damage in the areas requiring pesticide spraying.

[0053] In this embodiment, the collected multi-source data can also be denoised and standardized, and image fusion can be performed based on an image matching algorithm to obtain a real-world image to be planned for subsequent spraying path planning. The denoising method can be, for example, a Gaussian filtering algorithm; the image matching algorithm can be, for example, a SIFT algorithm; and the image fusion method can be, for example, wavelet transform fusion. These image processing techniques are well-known in the art and will not be elaborated further here. In other embodiments, preprocessing operations such as mechanical energy denoising and standardization can be performed only on the vegetation cover image of the Fritillaria thunbergii planting area to obtain the real-world image to be planned, providing a data foundation for subsequent planning of the UAV's flight path. It should be understood that in the field of UAV path planning technology, implementers can select appropriate images as the real-world images to be planned according to the specific planning scenario.

[0054] In other embodiments, denoising and region segmentation operations can be performed on the panoramic image of the Fritillaria thunbergii planting area to remove background interference and obtain an image containing only the Fritillaria thunbergii planting area. The denoising and region segmentation methods are well-known techniques in the field of image processing and will not be described in detail here.

[0055] Furthermore, in large-scale Fritillaria thunbergii cultivation scenarios, the spatial uniformity of vegetation cover is relatively strong due to the morphological characteristics of short plants (30-50cm in height) and high planting density, while the occurrence of pests and diseases exhibits significant spatial heterogeneity. The intensity, extent, and spread of pests and diseases vary significantly across different sub-regions. In large-scale Fritillaria thunbergii cultivation areas, multiple drones are required for multiple spraying operations. Since existing ant colony algorithms primarily focus on local optimization convergence for single drone paths, multiple drones may overlap in the same area during flight. Therefore, this step involves image recognition of the collected real-world images to be planned, analyzing the distribution of pest and disease damage at different locations within the cultivation area, and preliminarily determining the drone flight path based on the severity of damage using an ant colony algorithm.

[0056] Specifically, the drone's flight altitude is determined, and the Fritillaria thunbergii planting area corresponding to the planned real-world image is rasterized to obtain each raster region. The drone's flight distance from the centroid of each raster region to the drone's flight base station is obtained. The average pixel value of all pixels in the thermal image of each raster region is used as the pest and disease intensity of each raster region. The negative correlation coefficient of the drone's flight distance is used as the pheromone for the ant colony algorithm, and the pest and disease intensity is used as the priority weight. The ant colony algorithm is then used to plan the drone's flight path to obtain the spraying plan path for each drone.

[0057] More specifically, the negative correlation coefficient of the drone's flight distance can be obtained by using... The pheromone is acquired in a form that allows the ant colony algorithm to achieve higher base pheromones even when the drone's flight distance is shorter. This indicates a normalization process, such as the minimization normalization method. This indicates the distance the drone flew.

[0058] By recording the time the drone traverses each grid area along its spraying path, as well as the spatial coordinates of the centroid of each grid area, the three-dimensional spatial map can be expanded into a four-dimensional spatiotemporal map with spatial and temporal dimensions. It should be understood that all grid areas in each drone's spraying path are arranged in the order in which the drones traverse those areas.

[0059] In this implementation, the drone typically flies at an altitude 2-2.5 meters above the vegetation height in the grid area. The grid size can be set by the implementer based on the drone's flight altitude; the higher the drone's flight altitude, the larger the grid size. As a specific example, the product of the preset maximum grid size and the proportion of the drone's flight altitude is used as the grid size value. It should be understood that if the product is not an integer, it can be rounded down. In other embodiments, the implementer can set the drone's flight altitude and grid size according to the specific implementation scenario. For example, relevant personnel may divide the Fritillaria thunbergii planting area according to empirical dimensions based on the area to be planned.

[0060] In other embodiments, the drone flight distance for each grid area can be obtained by weighted summation of the vegetation height and the Euclidean distance between the centroid of each grid area and the drone base station. The larger the weight values ​​for vegetation height and Euclidean distance, the higher the attention given to that dimension of data. In this embodiment, the weights are set to the same value; in other embodiments, the implementer can set these values ​​according to the specific implementation scenario.

[0061] Step S200: Based on the distribution of the number of grid regions belonging to the UAV spraying planning path at the location of each grid region, and the pixel value of each grid region in the thermal imaging image, the grid regions are filtered to obtain conflict areas.

[0062] The core of conflict zone identification is to locate unreasonable areas where drones repeatedly spray water unnecessarily. Its main purpose is to distinguish between reasonable repetition and invalid conflict, quantify the degree of unreasonableness of the conflict, and provide precise adjustment targets for subsequent path reconstruction and reduction of repeated spraying costs.

[0063] Firstly, the essence of conflict zones lies in the mismatch between the degree of repeated spraying and the needs of pest and disease control. Excessive repeated spraying with mild pest and disease infestation constitutes ineffective conflict; conversely, excessive repeated spraying with severe pest and disease infestation indicates reasonable coordinated operation. A two-dimensional coupling judgment needs to be achieved through conflict characteristic indicators. Secondly, while conflict characteristic indicators quantify the degree of unreasonable repetition in each grid, not all high-ratio grids need to be included in the conflict zone (e.g., occasional repetition in individual grids). Threshold-based screening is necessary to identify high-frequency, concentrated ineffective conflict zones, avoiding inefficiency caused by scattered adjustments.

[0064] In this regard, such as Figure 2 As shown, the method for obtaining the conflict area can be implemented through steps S201 and S202.

[0065] Step S201: Obtain the proportion of the number of different UAVs whose locations are located in each grid area, and use the ratio of the proportion of the number to the pixel value of each grid area in the thermal imaging area as the conflict feature index of each grid area.

[0066] In large-area operation areas, multiple drones operating together repeatedly may cause spatiotemporal path conflicts due to repeated operations in the same area. This can lead to over-spraying or unreasonable spraying in some areas. Therefore, the main purpose of this step is to first obtain the proportion of paths for each grid cell, reflecting the density of repeated spraying; then obtain the pest and disease intensity corresponding to each grid cell, reflecting the control needs of the corresponding grid cell area; and finally, couple the results of the two-dimensional feature analysis to obtain conflict feature indicators, so as to quantify the conflict characteristics of each grid cell area in the drone's spraying planning path and avoid misjudgment from a single dimension.

[0067] Specifically, the percentage of each grid area that falls within the spraying paths of different drones refers to the proportion of the number of times a grid area is included in the spraying paths of different drones, relative to the total number of sprayings within the planting area. The percentage corresponding to each grid area reflects the density of repeated spraying of that grid area. When the same grid area exists in the spraying paths of multiple drones, it indicates that a spraying conflict has occurred between the flights of different drones.

[0068] It should be noted that during path planning, it is assumed that the agricultural drone spends an even amount of time passing through each grid area. Therefore, the lower the pest and disease intensity within a grid area, the more times the drone flies through it, indicating a greater likelihood of over-spraying. In this embodiment, for any given grid area, the average pixel value of all pixels in the thermal image of that grid area is taken as the pest and disease intensity of that area. The pest and disease intensity reflects the corresponding pixel value in the thermal image; a higher value indicates a greater pest and disease intensity within the grid area. In other embodiments, the sum of the pixel values ​​of all pixels within the grid area in the thermal image can also be used as the corresponding pixel value in the thermal image.

[0069] As a specific example, the ratio between the quantity percentage and the pest and disease intensity corresponding to each grid area can be used as a conflict characteristic index for each grid area.

[0070] It should be understood that the conflict characteristic index can also be the data after normalizing the calculation results of each grid area using the minimax normalization method. The normalization method is a well-known technique and is not restricted here.

[0071] Step S202: Based on the conflict characteristic index of each grid region, the grid regions are filtered to obtain conflict regions.

[0072] Specifically, grid areas with conflict characteristic indicators greater than or equal to a preset conflict threshold are designated as conflict areas. In this embodiment, the conflict threshold is set to 0.5. When the conflict characteristic indicator of a grid area is greater than or equal to the conflict threshold, it indicates that the repeated spraying by drones within that grid area is more severe, and the grid area is more likely to be an unreasonable area in the path planning process. Therefore, the grid area is identified as a conflict area, providing a priority adjustment target for subsequent path optimization, making path optimization more focused and efficient.

[0073] It should be understood that, within the Fritillaria thunbergii cultivation area, all grid areas other than conflict zones are considered non-conflict zones. It should be noted that the conflict threshold can be set by the implementer according to the specific implementation scenario. For example, by conducting experimental statistical analysis on historical flight data, the 95th percentile of the conflict characteristic indicators for all non-conflict zones can be set as the conflict threshold to ensure a high recall rate in non-conflict zones.

[0074] It should be understood that if a conflict area cannot be obtained, it means that the current path planning scheme of the drone is effective, and there is no conflict between different drones or the conflict is controllable. Therefore, no further path optimization is required.

[0075] Step S300: Based on the spatiotemporal distance distribution between each grid area and adjacent grid areas on the spraying planning path of each UAV, and combined with the pixel value of each conflict area in the thermal imaging image, the flight difficulty index of each grid area on the spraying planning path of each UAV is obtained.

[0076] The core of the flight difficulty index is to quantify the overall cost pressure of a drone flying in a specific grid. It mainly measures the basic flight cost of a single grid, identifies cost pain points in the path, integrates multi-dimensional features to judge the overall difficulty, and provides a quantitative basis for prioritizing the adjustment of high-difficulty grids when reconstructing the path.

[0077] Firstly, the planting areas of Fritillaria thunbergii are mostly sloping areas, and the difference in flight distance between adjacent grids directly affects the stability and energy consumption of drones. The greater the distance fluctuation, such as when suddenly entering a steep slope area from a flat area, the difference in distance between adjacent grids increases sharply. Drones need to frequently adjust their altitude and speed, and the higher the flight cost. It is necessary to capture this basic cost through the flight cost factor.

[0078] Secondly, the flight cost factor of a single grid needs to be measured in the cost system of the entire path. If the cost factor of a certain grid is high, its relative importance may be low if the average cost of the entire path is high. If only the cost of that grid is prominent, it is a cost pain point in the path, and subsequent reconstruction should prioritize adjustment. This relative importance needs to be quantified through spatiotemporal reconstruction parameters.

[0079] Thirdly, the difficulty of flight depends not only on the basic cost, but also on the needs of pest and disease control and the continuity of flight. The more severe the pest and disease in the grid, the more precise the flight is required to avoid missed spraying, which is more difficult. The greater the spatial and temporal distance between adjacent grids, the more dispersed the flight is, the more room the drone can adjust, and the lower the difficulty. It is necessary to achieve comprehensive quantification through the coupling of these three factors.

[0080] In this regard, such as Figure 3 As shown, the method for obtaining the flight difficulty index can be implemented by steps S301 to S303.

[0081] Step S301: Based on the flight distance distribution between the selected grid area and adjacent grid areas on each UAV spraying planning path, obtain the flight cost factor of the selected grid area on each UAV spraying planning path.

[0082] Specifically, this embodiment uses the spraying planning path of any drone as an example for illustration, and any grid area on the spraying planning path is recorded as the selected grid area. The main purpose of this step is to calculate the difference in flight distance between the current flight position of the drone in each grid area and the positions before and after it in the path, based on the characteristic that the slope of Fritillaria thunbergii planting in different areas of sloping land is significantly different, and to quantify the flight cost data of each grid dimension in the drone's spraying flight path.

[0083] More specifically, the first step is to obtain the drone flight distance for each grid area along the planned spraying path.

[0084] It should be understood that the method for obtaining the drone's flight distance is described in step S100, and will not be repeated here.

[0085] The second step is to obtain the absolute value of the difference in drone flight distance between the selected grid area and the first adjacent grid area to obtain the first flight difference; and to obtain the absolute value of the difference in drone flight distance between the selected grid area and the second adjacent grid area to obtain the second flight difference; the sum of the first flight difference and the second flight difference is the flight cost factor of the selected grid area; wherein, the first adjacent grid area is the previous grid area adjacent to the selected grid area on the spraying planning path, and the second adjacent grid area is the next grid area adjacent to the selected grid area on the spraying planning path.

[0086] First flight difference This represents the difference in flight distance between the i-th drone and the x-th grid area in the spraying planning path, specifically the second flight difference. This represents the difference in flight distance between the i-th drone and the x-th (x+1)-th grid area within the planned spraying path. This represents the flight distance of the i-th drone in the x-th grid area of ​​the spraying plan path. This represents the flight distance of the i-th drone in the (x-1)-th grid area of ​​the spraying plan path. This represents the drone's flight distance in the (x+1)th grid area of ​​the spraying planning path of the i-th drone.

[0087] In the drone spraying path planning, the difference in flight distance between two adjacent grid areas reflects the cost of the current flight path of the i-th drone. The larger the distance difference, the longer the route, indicating more drastic changes in distance at that location, higher energy and time costs for the drone, and a greater likelihood of path adjustment. The flight cost factor mainly focuses on the local path cost of a single drone in a certain grid area, quantifying the degree of fluctuation in flight distance before and after that grid area location. The greater the distance fluctuation, the higher the flight cost, and the more priority should be given to path adjustment, providing a data foundation for subsequent multi-drone coordination.

[0088] It should be noted that if the selected grid is the first grid (without a first adjacent grid) or the last grid (without a second adjacent grid) of the path, only the corresponding second flight difference or first flight difference is calculated to obtain the flight cost factor.

[0089] Step S302: The sum of the differences between the flight cost factor of the selected grid area on the selected spraying planning path and the flight cost factor of the other UAV spraying planning path is used as the cost fluctuation degree of the selected grid area. The ratio of the flight cost factor of the selected grid area on the selected spraying planning path to the cost fluctuation degree is used as the spatiotemporal reconstruction parameter of the selected grid area on the selected spraying planning path.

[0090] The main purpose of this step is to expand from the local cost of a single drone to the global coordination of multiple drones, quantify the relative cost pressure of a drone in each grid area, that is, whether the cost of a drone in a grid area is significantly higher than that of other drones, and provide a basis for prioritizing the multi-drone path adjustment.

[0091] In this embodiment, the spraying planning path of the i-th UAV is taken as the selected spraying planning path, and the x-th grid region of the selected spraying planning path is taken as the selected grid region. As a first specific example, the method for obtaining the spatiotemporal reconstruction parameters of the selected grid region in the selected spraying planning path can be expressed as follows:

[0092]

[0093] in, This represents the spatiotemporal reconstruction parameters of the x-th grid region in the spraying planning path of the i-th drone. This represents the flight cost factor for the x-th grid region in the spraying planning path of the i-th drone. This represents the flight cost factor of the x-th grid area in the spraying planning path of the n-th drone. This indicates the total number of drones whose spraying plans pass through the selected grid area.

[0094] The sum of the differences between the flight cost factors of the selected grid area on the selected spraying planning path and the flight cost factors of other UAV spraying planning paths is the degree of cost fluctuation. It reflects the differences between the flight cost factors of all UAVs during the actual flight of the UAVs, and thus reflects the cost fluctuation of the selected grid area in all conflicting planning paths.

[0095] When cost fluctuations are significant, it indicates substantial differences in flight costs among different drones and high volatility in conflict zones, potentially increasing the complexity of path reconstruction. In this case, smaller values ​​for the spatiotemporal reconstruction parameters mean that the current drone has a lower priority for path reconstruction within the selected grid area, as cost fluctuations make adjustments more difficult.

[0096] When the cost fluctuation is small, the flight cost fluctuates little among different UAVs, and the spatiotemporal reconstruction parameter has a larger value. This indicates that the flight cost of the current UAV in the selected grid area is relatively prominent, and the priority of path reconstruction is higher. This means that the greater the flight route consumption of the current i-th UAV in the selected grid area, the easier it is to reconstruct and adjust the UAV path in the selected grid area.

[0097] The spatiotemporal reconstruction parameters reflect the relative relationship between the current UAV flight cost and regional cost fluctuations within the selected grid area. A larger spatiotemporal reconstruction parameter indicates greater relative cost pressure for the UAV within the selected grid area, higher feasibility of path reconstruction, a more significant reduction in global cost after adjustment, and a higher adjustment priority.

[0098] It should be noted that when the selected grid area does not exist in the spraying flight path of other drones, the cost fluctuation level cannot be obtained. In this case, the cost fluctuation level corresponding to the grid area can be directly set to 1, which means that when only one drone passes through the grid area, the priority of its path reconstruction is directly determined by its flight cost factor.

[0099] Step S303: Based on the spatiotemporal distance distribution between the selected grid area and adjacent grid areas on the spraying planning path, the pixel values ​​in the thermal imaging image, and the spatiotemporal reconstruction parameters, the flight difficulty index of the selected grid area is obtained.

[0100] The main purpose of this step is to incorporate the needs of pest and disease control and the spatiotemporal characteristics of the route, and to quantify the actual flight difficulty of the drone at each grid position. The lower the difficulty, the more suitable it is to adjust the path; the higher the difficulty, the more necessary it is to maintain the original path to avoid missing areas with severe pests and diseases.

[0101] Specifically, the first step is to determine the spatiotemporal distance between the selected grid area and the second adjacent grid area based on the time corresponding to the selected grid area on the spraying planning path and the location of the selected grid area.

[0102] As a concrete example, the three-dimensional coordinates of a selected grid area on the selected spraying planning path, combined with time, constitute the spatiotemporal coordinates of the selected grid area. Using the same method, the spatiotemporal coordinates of the next adjacent grid area (the second adjacent grid area) on the selected spraying planning path can be obtained. Then, based on the spatiotemporal coordinates of the selected grid area and the second adjacent grid area, the Euclidean distance is calculated as the spatiotemporal distance between the selected grid area and the second adjacent grid area. This reflects the concentration of routes within the selected grid area on the selected spraying planning path; a larger spatiotemporal distance indicates more dispersed routes, and a smaller spatiotemporal distance indicates denser routes. It should be understood that the selected spraying planning path refers to the spraying planning path of any single UAV.

[0103] The second step is to use the ratio between the pixel value of the selected grid region in the thermal imaging image and the product of the spatiotemporal distance and the spatiotemporal reconstruction parameter as the flight difficulty index of the selected grid region.

[0104] As a concrete example, the method for obtaining the flight difficulty index of a selected grid area along a selected spraying path can be expressed by the formula:

[0105]

[0106] in, This represents the flight difficulty index of the selected grid area on the selected spraying planning path, where i represents the spraying planning path of the i-th UAV and x represents the x-th grid area on the spraying planning path of the i-th UAV. This represents the pixel value of the selected grid area in the thermal image. It can be the average or cumulative value of the pixel values ​​of all pixels in the grid area, reflecting the severity of pests and diseases in the selected grid area. This represents the spatiotemporal reconstruction parameters of the x-th grid region in the spraying planning path of the i-th drone. This indicates the spatiotemporal distance between the selected raster region and the second adjacent raster region. Indicates the selection of a grid area. This indicates the second adjacent grid area.

[0107] Spacetime Distance The larger the distance, the greater the spatial and temporal interval between different spraying areas in the current flight path. This indicates that the flight path is relatively unconcentrated, with more redundant space for adjustment and relatively lower flight difficulty. During flight, route replanning can be used to reduce interference from collisions or flight conflicts, thus reducing flight difficulty. This suggests that there should be a negative correlation between spatial and temporal distance and flight difficulty, meaning that as spatial and temporal distance increases, flight difficulty decreases.

[0108] The larger the value, the greater the relative cost pressure on the drone within that grid, the more it needs to reduce costs through path reconstruction, and the lower its flight difficulty. The higher the value, the stronger the pest and disease infestation, requiring precise spraying. Theoretically, the greater the difficulty in adjustment, which means the flight becomes more difficult. This indicates that areas with severe pest and disease infestations need to be handled more carefully, which may increase the difficulty of flight.

[0109] It should be noted that for selected grid areas where the adjacent next grid cannot be obtained, the adjacent previous grid area can be used in the calculation process of the flight difficulty index. That is, for the tail grid, the spatiotemporal distance between the adjacent previous grid area and the grid area is used to calculate the flight difficulty index.

[0110] The flight difficulty index is a comprehensive quantification of objective limitations and the necessity of adjustments. The smaller the value, the easier it is to adjust the path and the lower the flight difficulty.

[0111] Step S400: Based on the changes in the flight difficulty index of each conflict area and the flight difficulty index of other grid areas on the spraying planning path of each drone, the spraying planning path of the drone is replanned.

[0112] The core of path reconstruction is to replace high-difficulty conflict areas with low-difficulty, non-conflict areas, while ensuring the continuity and integrity of the reconstructed path. This is mainly achieved by screening effective alternative areas, judging the rationality of alternative solutions, ensuring the continuity of the reconstructed path, and providing key adjustment solutions for ultimately achieving a non-repetitive, low-cost spraying path.

[0113] Firstly, the core issues in conflict areas are repeated spraying and high flight difficulty. Reconstruction requires first screening for conflict-free and replaceable candidate grids, then quantifying the replacement value through difficulty difference to ensure that the path difficulty is reduced after replacement. Secondly, not all candidate grids can meet the requirement of significantly reducing difficulty; the optimal adjustment target needs to be selected through threshold screening. Furthermore, due to path continuity, subsequent grids need to be replanned after replacing conflict areas to avoid new cost issues. Thirdly, adjustments to individual conflict areas need to be integrated into the global logic of the entire path to avoid local adjustments leading to new conflicts. Finally, a complete reconstructed path is obtained by integrating the reconstruction parameters of all candidate grids.

[0114] In this regard, such as Figure 4 As shown, the sub-steps of step S400 can be implemented by steps S401 to S403.

[0115] Step S401: For any spraying planning path, obtain other non-conflict areas besides the target conflict area as candidate grid areas; the target conflict area is any conflict area on the spraying planning path.

[0116] Non-conflict areas do not have the problem of repeated spraying and are already covered in the initial plan, ensuring that no areas that need spraying are missed after replacement (such as areas infested with pests and diseases of Fritillaria thunbergii), and avoiding blind spots in prevention and control due to reconstruction. Conflict areas have the greatest impact on the overall operation cost (repeated spraying) and coverage integrity (potential blind spots). Prioritizing these areas can achieve global optimization with minimal adjustment costs.

[0117] Step S402: Based on the difference in flight difficulty index between the target conflict area and each candidate grid area, obtain the path reconstruction parameters for each candidate grid area.

[0118] Specifically, the difference in flight difficulty index between the target conflict area and each candidate grid area is used as the path reconstruction parameter for each candidate grid area.

[0119] The main purpose of this step is to analyze the changes in the difficulty of route flight reconstruction before and after a single drone moves from a selected grid position to each candidate grid position, taking into account the simultaneous operation of multiple agricultural drones. This quantifies the impact of route reconstruction on drone route reconstruction, meaning that the path reconstruction parameters reflect the changes in the difficulty of flight reconstruction before and after the change in flight position.

[0120] It should be noted that conflict areas are the concentrated outbreak points of path conflicts. Optimizing paths in these areas can significantly reduce computational load, making path optimization more focused and efficient. The larger the difference, the larger the value of the path reconstruction parameter, indicating that replacing the conflict area with this candidate significantly reduces flight difficulty; the smaller the difference, the smaller the value of the path reconstruction parameter, indicating that the replacement has a limited effect on improving difficulty, thus accurately quantifying the value of the alternative.

[0121] Step S403: Adjust the spraying planning path according to the path reconstruction parameters of each candidate grid area to obtain the reconstructed path of the UAV spraying planning path.

[0122] Specifically, candidate grid areas corresponding to the path reconstruction parameters that are greater than the preset reconstruction threshold and the maximum value of the path reconstruction parameters are selected as adjustment grid areas; target conflict areas in the spraying planning path are adjusted to adjustment grid areas, and the grid areas after the target conflict areas in the adjusted spraying planning path are replanned to obtain the reconstructed path.

[0123] In this embodiment, the reconstruction threshold is set to 0. That is, when the path reconstruction parameter of each candidate grid area is less than 0, it means that when the target conflict area is adjusted by the candidate grid area for path optimization, the flight difficulty increases after the flight route is replanned. This indicates that the effect of adjusting the current route by the candidate grid area is poor, and it is not necessary to optimize the local route by the candidate grid area.

[0124] When the path reconstruction parameter of each candidate grid region is greater than 0, it indicates that when the target conflict area is adjusted using the candidate grid region for path optimization, the flight difficulty of the UAV is reduced after route reconstruction. This indicates that the effect of adjusting the current route using the candidate grid region is good, and the candidate grid region can be used for local route optimization.

[0125] The specific optimization operation is to adjust the target conflict area in the spray planning path to an adjustment grid area. For example, if the target conflict area is the m-th grid area in the spray planning path, the adjustment grid area is taken as the m-th grid area in the spray planning path, and the m-th grid area in the original spray planning path and all subsequent planning paths are replanned to obtain the optimized and adjusted reconstructed path.

[0126] After replacing the target conflict area in the path with an adjustment grid, all subsequent grids need to be replanned. This is because the Zhejiang fritillary bulb planting area is mostly on slopes, and the flight difficulty of the grids in the path is strongly correlated with the spatiotemporal distance between adjacent grids. If the distance between an adjustment grid and the next grid is significantly different from the distance between the original conflict area and the next grid, the flight difficulty of subsequent grids may increase sharply. If a sudden change in distance occurs on a steep slope, it will lead to new path bottlenecks.

[0127] The path optimization strategy of this invention ensures the mission integrity of each UAV while reducing global conflicts through individual optimization and collaborative verification, perfectly adapting to the needs of multi-UAV collaborative spraying in Fritillaria thunbergii cultivation areas. The reconfiguration decision incorporates global information from multi-UAV collaboration, ultimately achieving global operation optimization through local path adjustments of each UAV.

[0128] After optimizing and adjusting the spraying planning path for each drone, in other embodiments, it can be verified whether there are new conflict areas in the reconstructed path after optimization and adjustment. If so, iterative optimization can be performed according to the steps of the embodiments of the present invention.

[0129] The essence of replanning is to dynamically adapt to environmental changes, aiming for global optimization and finding a balance between minimizing conflicts and ensuring complete coverage. Its logic chain can be simplified as follows: generating a spraying planning path, detecting new conflict areas, calculating and adjusting parameters, and globally optimizing the path. Each step is constrained by the characteristics of Fritillaria thunbergii cultivation (small size, sloping terrain, and susceptibility to pests and diseases) and the limitations of drone operations (short flight time, multi-drone collaboration), ultimately achieving the goals of reducing repeated spraying, avoiding blind spots, and reducing costs.

[0130] It should be noted that in practical applications, the spraying progress can be monitored in real time using the infrared sensors mounted on the drone. When changes in the severity of pests or diseases or changes in terrain are detected in the affected area, a new conflict analysis can be performed to determine whether there are conflict areas and whether path optimization is necessary.

[0131] In summary, the embodiments of the present invention identify conflict points on multiple paths during flight, adjust the conflict based on the degree of pests and diseases at the conflict points, and optimize the path planning of drones based on different single-operation distances of drones and the complexity of regional terrain, thereby reducing the cost and time of repeated spraying operations.

[0132] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application 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 of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for intelligent planning of a pesticide spraying path for preventing and treating diseases and pests of Fritillaria thunbergii Miq., characterized in that, The method includes the following steps: Acquire thermal imaging images of the Fritillaria thunbergii planting area and the spraying planning path of each UAV, wherein the spraying planning path includes several grid areas of the Fritillaria thunbergii planting area; Based on the distribution of the number of grid regions belonging to the drone spraying planning path at the location of each grid region, and the pixel value of each grid region in the thermal imaging image, the grid regions are filtered to obtain conflict areas. Based on the spatiotemporal distance distribution between each grid area and adjacent grid areas on the spraying planning path of each UAV, and combined with the pixel value of each conflict area in the thermal imaging image, the flight difficulty index of each grid area on the spraying planning path of each UAV is obtained. Based on the changes in the flight difficulty index of each conflict area and the flight difficulty index of other grid areas on the spraying planning path of each drone, the spraying planning path of the drone is replanned. The method for obtaining the flight difficulty index includes: Based on the flight distance distribution between the selected grid area and adjacent grid areas on each drone spraying planning path, the flight cost factor of the selected grid area on each drone spraying planning path is obtained. The sum of the differences between the flight cost factor of the selected grid area on the selected spraying planning path and the flight cost factor of the other UAV spraying planning paths is used as the cost fluctuation degree of the selected grid area. The ratio of the flight cost factor of the selected grid area on the selected spraying planning path to the cost fluctuation degree is used as the spatiotemporal reconstruction parameter of the selected grid area on the selected spraying planning path. The selected grid area is any grid area on the spray planning path; the selected spray planning path is any spray planning path of any UAV. Based on the spatiotemporal distance distribution between the selected grid area and adjacent grid areas on the spraying planning path, the pixel values ​​in the thermal imaging image, and the spatiotemporal reconstruction parameters, the flight difficulty index of the selected grid area is obtained.

2. The intelligent planning method for spraying pesticide paths for pest and disease control of Fritillaria thunbergii as described in claim 1, characterized in that, The process of filtering grid regions to identify conflict areas based on the distribution of the number of grid regions belonging to the UAV spraying plan path at each grid region's location and the pixel value of each grid region in the thermal imaging image specifically includes: The proportion of the number of spraying planning paths of different drones at the location of each grid area is obtained, and the ratio of the proportion of the number to the pixel value of each grid area in the thermal imaging area is used as the conflict feature index of each grid area. Based on the conflict characteristic index of each grid region, the grid regions are filtered to obtain conflict areas.

3. The intelligent planning method for spraying pesticide paths for pest and disease control of Fritillaria thunbergii according to claim 2, characterized in that, The process of filtering grid regions based on their conflict characteristic indicators to obtain conflict regions specifically includes: The grid area corresponding to the conflict feature index being greater than or equal to the preset conflict threshold is defined as the conflict area.

4. The intelligent planning method for spraying pesticide paths for pest and disease control of Fritillaria thunbergii as described in claim 1, characterized in that, The method of obtaining the flight cost factor of the selected grid area on each UAV spraying planning path based on the flight distance distribution between the selected grid area and adjacent grid areas on each UAV spraying planning path specifically includes: Obtain the drone flight distance for each grid area along the spraying plan path; For any drone's spraying planning path, the absolute value of the difference between the drone's flight distance between the selected grid area and the first adjacent grid area is obtained as the first flight difference; the absolute value of the difference between the drone's flight distance between the selected grid area and the second adjacent grid area is obtained as the second flight difference; the sum of the first flight difference and the second flight difference is the flight cost factor of the selected grid area. The first adjacent grid area is the preceding grid area adjacent to the selected grid area on the spraying planning path, and the second adjacent grid area is the following grid area adjacent to the selected grid area on the spraying planning path.

5. The intelligent planning method for spraying pesticides to control pests and diseases of Fritillaria thunbergii according to claim 4, characterized in that, The process of obtaining the flight difficulty index of the selected grid area based on the spatiotemporal distance distribution between the selected grid area and adjacent grid areas on the spraying planning path, the pixel values ​​in the thermal imaging image, and the spatiotemporal reconstruction parameters specifically includes: Based on the time corresponding to the selected grid area on the spraying planning path and the location of the selected grid area, the spatiotemporal distance between the selected grid area and the second adjacent grid area is determined. The ratio between the pixel value of the selected grid region in the thermal imaging image and the product of the spatiotemporal distance and the spatiotemporal reconstruction parameter is used as the flight difficulty index of the selected grid region.

6. The intelligent planning method for spraying pesticide paths for pest and disease control of Fritillaria thunbergii according to claim 1, characterized in that, The process involves replanning the drone spraying path based on changes in the flight difficulty index of each conflict zone and other grid zones along the spraying path of each drone. This includes: For any given spraying plan path, obtain the non-conflict areas other than the target conflict area as candidate grid areas; the target conflict area is any conflict area on the spraying plan path. Based on the difference in flight difficulty index between the target conflict area and each candidate grid area, the path reconstruction parameters for each candidate grid area are obtained; Based on the path reconstruction parameters of each candidate grid area, the spraying planning path is adjusted to obtain the reconstructed path of the UAV spraying planning path.

7. The intelligent planning method for spraying pesticide paths for pest and disease control of Fritillaria thunbergii according to claim 6, characterized in that, The method of obtaining path reconstruction parameters for each candidate grid region based on the difference in flight difficulty index between the target conflict region and each candidate grid region specifically includes: The difference in flight difficulty index between the target conflict area and each candidate grid area is used as the path reconstruction parameter for each candidate grid area.

8. The intelligent planning method for spraying pesticide paths for pest and disease control of Fritillaria thunbergii according to claim 6, characterized in that, The step of adjusting the spraying planning path based on the path reconstruction parameters of each candidate grid region to obtain the reconstructed path of the UAV spraying planning path specifically includes: The candidate raster region corresponding to the path reconstruction parameter being greater than the preset reconstruction threshold and the maximum value of the path reconstruction parameter is used as the adjustment raster region. The target conflict area in the spraying planning path is adjusted to an adjustment grid area. The grid area after the target conflict area in the adjusted spraying planning path is then replanned to obtain the reconstructed path.

9. The intelligent planning method for spraying pesticide paths for pest and disease control of Fritillaria thunbergii according to claim 1, characterized in that, The method for obtaining the spraying planning path for each drone specifically includes: The drone flight distance for each grid area is obtained, and the average pixel value of all pixels in the thermal image of each grid area is used as the pest and disease intensity of each grid area. The negative correlation coefficient of the drone's flight distance is used as the pheromone for the ant colony algorithm, and the intensity of the pests and diseases is used as the priority weight. The ant colony algorithm is used to plan the flight path of the drones to obtain the spraying plan path for each drone.

Images