Crop disease and pest image acquisition point planning method based on unmanned aerial vehicle

Abstract

The invention provides a crop disease and insect pest image acquisition point planning method based on an unmanned aerial vehicle, and the method comprises the steps: obtaining image data of a crop planting region through the unmanned aerial vehicle or a satellite at different time nodes in a specified time period, constructing an image data set based on a time sequence, and carrying out the preprocessing; inputting the processed image data set into a growth vigor analysis model for processing to obtain crop growth vigor information, analyzing crops with abnormal growth vigor according to the crop growth vigor information and determining positions, planning an unmanned aerial vehicle route according to the positions of the crops with abnormal growth vigor, and controlling the unmanned aerial vehicle to advance according to the unmanned aerial vehicle route. High-resolution images of corresponding crops with abnormal growth vigor are collected when the high-resolution images pass through collection sites in the air route, the high-resolution images are input into a pest and disease damage analysis model to be processed, pest and disease damage identification results are obtained, and pest and disease damage prediction information and prevention and control suggestions are generated according to the pest and disease damage identification results and output. And a user is assisted to take reasonable pest control measures in time.

Description

Technical Field

[0001] The present invention relates to the technical field of pest and disease monitoring, and in particular to a method for planning crop pest and disease image acquisition points based on an unmanned aerial vehicle. Background Art

[0002] With the advancement of agricultural modernization, large-scale planting has become the main trend of global agricultural production. However, large-scale planting also brings new challenges, such as the rapid spread of pests and diseases, and the increased complexity of monitoring and management. Pests and diseases are one of the main factors affecting crop growth and yield. They directly attack the roots, stems, leaves, flowers, fruits and other parts of crops, interfere with the normal physiological functions of crops, and cause crop growth retardation, yield reduction or even crop failure. Therefore, how to efficiently and accurately monitor and manage pests and diseases has become a key issue that needs to be solved in large-scale planting.

[0003] Timely and accurate monitoring of crop pests and diseases has important economic and ecological significance. Traditional pest and disease monitoring mainly relies on manual surveys. Farmers or technicians regularly patrol the planting areas and detect the occurrence of pests and diseases through visual observation or simple tools. Although this method is intuitive, it has problems such as low efficiency, limited coverage, and strong subjectivity, and it is difficult to meet the needs of large-scale planting. In addition, manual monitoring is time-consuming and labor-intensive, and it is difficult to implement in high-density planting areas or complex terrain areas.

[0004] In recent years, the application of drone technology in the agricultural field has gradually become popular, providing a new solution for crop pest and disease monitoring. Drones are equipped with sensors such as multispectral cameras and thermal imagers, which can quickly obtain high-resolution image data of large-scale farmland. By analyzing the image data, the characteristics of crop pests and diseases, such as leaf discoloration, wilting, and traces of pests, can be identified. In addition, drones can also be combined with artificial intelligence algorithms to realize automatic identification and classification of pests and diseases, significantly improving monitoring efficiency and accuracy.

[0005] Although drone technology has shown great potential in crop pest and disease monitoring, it still faces many challenges in its practical application in large-scale planting areas. First, large-scale planting areas are vast, and drones need to cover a large area, so flight path planning and battery life become key limiting factors. Secondly, for large-scale planting areas, the amount of data collected by drones will also increase accordingly, which puts higher requirements on the computing resources used to process data while ensuring timeliness. Summary of the invention

[0006] The object of the present invention is to provide a method for planning image acquisition points of crop pests and diseases based on an unmanned aerial vehicle (UAV), which analyzes crops with abnormal growth in a planting area based on RGB images and multispectral remote sensing images, determines the positions of the crops with abnormal growth, determines the image acquisition points of pests and diseases based on the positions of the crops with abnormal growth, and plans the UAV flight path, so as to improve the efficiency of image acquisition of pests and diseases while reducing the energy consumption of the UAV.

[0007] To achieve the above object of the invention, the present invention provides a method for planning image acquisition points of crop pests and diseases based on an unmanned aerial vehicle (UAV), the method comprising:

[0008] Obtaining image data of a crop planting area at different time nodes within a specified time period by using a UAV or a satellite, and constructing a time-series-based image data set;

[0009] Preprocessing the image data set, inputting the processed image data set into a growth analysis model for processing, and obtaining crop growth information, where the crop growth information is used to characterize the growth of crops at different positions in the crop planting area;

[0010] Analyzing the crops with abnormal growth according to the crop growth information, and determining the positions of the crops with abnormal growth;

[0011] Planning a UAV flight path according to the positions of the crops with abnormal growth, where the UAV flight path includes a number of acquisition points, and each acquisition point corresponds to a crop with abnormal growth;

[0012] Controlling the UAV to travel along the UAV flight path, and collecting high-resolution images of the corresponding crops with abnormal growth when passing through the acquisition points;

[0013] Inputting the high-resolution images of the crops with abnormal growth into a pest and disease analysis model for processing, obtaining pest and disease identification results, and generating and outputting pest and disease prediction and prevention advice according to the pest and disease identification results.

[0014] Furthermore, the image data includes RGB images and multispectral remote sensing images.

[0015] Furthermore, preprocessing the image data set specifically includes the following operations:

[0016] Stitching the image data in the image data set into an orthoimage;

[0017] Performing geometric correction and radiometric correction on the image data in the image data set;

[0018] Performing band synthesis on multispectral remote sensing images of different bands;

[0019] Removing noise from the image data and cropping the image data.

[0020] Further, the radiation correction specifically includes the following operations:

[0021] Collect multispectral remote sensing images of the crop planting area through the first multispectral camera, and collect downwelling light data of the same moment and the same area through the downwelling light sensor;

[0022] Collect reference board data of the same area through the second multispectral camera while the first multispectral camera is taking pictures, and collect reference board data at regular intervals to obtain a set of reference board data based on time series;

[0023] Perform radiation correction on the multispectral remote sensing images collected by the first multispectral camera based on the set of reference board data and the downwelling light data.

[0024] Further, input the processed image data set into the growth analysis model for processing to obtain crop growth information, which specifically includes the following operations:

[0025] Construct a color space model based on the RGB images in the image data set, and calculate the vegetation index of the corresponding crop planting area according to the color space model;

[0026] Extract texture features from the multispectral remote sensing images in the image data set;

[0027] Use the obtained vegetation index and texture features as feature parameters, screen the feature parameters to obtain the final feature parameters, and construct a growth analysis model based on the final feature parameters.

[0028] Further, screening the feature parameters specifically includes the following operations:

[0029] Perform dimensionless processing on the feature parameters;

[0030] Analyze the correlation between all feature parameters;

[0031] Construct a growth evaluation index system for crops, calculate the correlation between each feature parameter and the growth evaluation index, screen out several feature parameters with relatively weak correlation, and obtain the final feature parameters;

[0032] Calculate the geodesic distance of the mutually correlated final feature parameters by the equidistant feature mapping method, construct a geodesic distance matrix, and perform multi-dimensional scaling transformation on the geodesic distance matrix to obtain a low-dimensional feature space.

[0033] Further, plan the UAV flight path according to the location of the crops with abnormal growth, which specifically includes the following operations:

[0034] Calculate the representativeness of the image data corresponding to the positions of each crop with abnormal growth, where the representativeness is the ratio of the characteristic parameters when the crop growth is abnormal to the standard characteristic parameters when the crop growth is normal;

[0035] Cluster the positions of each crop with abnormal growth according to the representativeness, obtain multiple clustering clusters, and screen out the representative sites of each clustering cluster;

[0036] Plan the UAV flight path based on all the representative sites.

[0037] Further, planning the UAV flight path based on all the representative sites specifically includes the following operations:

[0038] Convert the longitude and latitude coordinates of all the representative sites into plane coordinates, and calculate the distance matrix between all the representative sites;

[0039] Take the distance matrix as the input, calculate the shortest path passing through all the representative sites through a heuristic algorithm, and take the finally calculated shortest path as the UAV flight path.

[0040] Further, after calculating the shortest path passing through all the representative sites through a heuristic algorithm, the following operations are also included:

[0041] Construct a flight energy consumption model for the UAV;

[0042] Aim at the lowest comprehensive energy consumption, and optimize and adjust the shortest path based on the real-time wind direction and wind speed and the flight energy consumption model to obtain the final UAV flight path.

[0043] Compared with the prior art, the beneficial effects of the present invention are:

[0044] 1. The present invention constructs a time-series image data set of the crop planting area, inputs it into the growth analysis model to analyze the crop growth information, determines the crops with abnormal growth and their positions, and plans the UAV flight path according to the positions of the crops with abnormal growth, so that the UAV can collect high-resolution images of the crops with abnormal growth based on the collection sites passed by the UAV flight path to analyze whether they are damaged by pests and diseases, which can significantly improve the data collection efficiency and reduce the UAV energy consumption in large-scale planting areas;

[0045] 2. The present invention analyzes and processes the high-resolution images of the crops with abnormal growth through the pest and disease analysis model, generates pest and disease prediction and forecast information and prevention and control suggestions according to the analysis results and outputs them, which can help farmers or technicians quickly determine the pest and disease situation and take reasonable prevention and control measures to prevent further damage. Description of the Drawings

[0046] To more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the accompanying drawings required for the description of the embodiments. Obviously, the accompanying drawings in the following description are only the preferred embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings.

[0047] Figure 1 It is a schematic diagram of the overall process of a method for planning crop pest and disease image acquisition points based on an unmanned aerial vehicle provided by an embodiment of the present invention. Detailed implementation manners

[0048] The following describes the principles and features of the present invention with reference to the accompanying drawings. The listed embodiments are only used to explain the present invention and are not intended to limit the scope of the present invention.

[0049] Referring to Figure 1 , this embodiment provides a method for planning crop pest and disease image acquisition points based on an unmanned aerial vehicle. The method includes:

[0050] S101. Obtain the image data of the crop planting area at different time nodes within a specified time period through an unmanned aerial vehicle or satellite, and construct a time-series-based image data set.

[0051] In this embodiment, the image data includes RGB images and multispectral remote sensing images. The time-series-based image data set contains the image data of the crop planting area corresponding to different time nodes within the specified time period.

[0052] S102. Preprocess the image data set, input the processed image data set into a growth analysis model for processing, and obtain crop growth information, which is used to characterize the crop growth of different positions in the crop planting area.

[0053] S103. Analyze the crops with abnormal growth according to the crop growth information, and determine the positions of the crops with abnormal growth.

[0054] S104. Plan the unmanned aerial vehicle flight path according to the positions of the crops with abnormal growth. The unmanned aerial vehicle flight path includes several acquisition points, and each acquisition point corresponds to a crop with abnormal growth.

[0055] S105. Control the unmanned aerial vehicle to travel according to the unmanned aerial vehicle flight path. When passing through the acquisition point, acquire the high-resolution images of the corresponding crops with abnormal growth.

[0056] S106. Input the high-resolution images of the crops with abnormal growth into a pest and disease analysis model for processing, obtain the pest and disease identification results, generate pest and disease prediction and forecast information and prevention and control suggestions according to the pest and disease identification results, and output them.

[0057] In this step, the pest and disease analysis model is obtained by pre-training a standard neural network model with labeled sample crop images, sample pest and disease identification results, sample pest and disease prediction and forecasting information, and sample control suggestions.

[0058] For the method provided in this embodiment, on the one hand, by collecting image data at different time points within a specified time period, after preprocessing the image data, it is input into the growth analysis model for processing to obtain crop growth information, so as to further determine the positions of crops with abnormal growth, and plan the UAV flight path according to the positions of crops with abnormal growth. When the UAV moves along the UAV flight path, it will pass through each collection point. When the UAV reaches the collection point, high-resolution images of the crops with abnormal growth at the collection point are collected for subsequent specific analysis of whether the corresponding crops are affected by pests and diseases. The method can effectively improve the efficiency of pest and disease control and reduce the energy consumption of UAV flight collection by specifically collecting images of crops with abnormal growth to analyze whether they are affected by pests and diseases.

[0059] The method provided in this embodiment also inputs the high-resolution images of the crops with abnormal growth into the pest and disease analysis model for processing to obtain pest and disease identification results, and generates pest and disease prediction and forecasting information and control suggestions according to the pest and disease identification results and outputs them, so as to help farmers or technicians quickly determine the pest and disease conditions of the crops with abnormal growth and the corresponding control measures.

[0060] As a possible implementation manner, preprocessing the image data set specifically includes the following operations:

[0061] S201. Stitch the image data in the image data set into an orthoimage.

[0062] In this implementation manner, the purpose of stitching the image data into an orthoimage is to eliminate geometric distortion and perspective error and generate a two-dimensional image with a unified scale and high geometric accuracy.

[0063] S202. Perform geometric correction and radiometric correction on the image data in the image data set.

[0064] In an outdoor environment, changes in lighting conditions are inevitable. And optical sensors are very sensitive to changes in lighting conditions, which may cause the image data collected by the optical sensors to be inaccurate. This implementation manner eliminates the influence of changes in external environmental lighting conditions on the accuracy of image data through radiometric correction.

[0065] S203. Synthesize multi-spectral remote sensing images of different bands.

[0066] S204. Remove the noise in the image data and crop the image data.

[0067] In this embodiment, the image data is cropped to eliminate the deformed and abnormal edge data generated during splicing, avoiding its impact on subsequent analysis.

[0068] As a further possible embodiment, the radiation correction includes the following operations:

[0069] S301. Collect the multi-spectral remote sensing image of the crop planting area through the first multi-spectral camera, and collect the downwelling light data of the same moment and the same area through the downwelling light sensor.

[0070] S302. While the first multi-spectral camera is taking pictures, collect the reference panel data of the same area through the second multi-spectral camera, and collect the reference panel data every once in a while to obtain a set of reference panel data based on time series.

[0071] The set of reference panel data consists of the reference panel data collected by the second multi-spectral camera for the first time (i.e., the reference panel data of the same area collected while the first multi-spectral camera is taking pictures) and the reference panel data collected at subsequent intervals.

[0072] S303. Based on the set of reference panel data and the downwelling light data, perform radiation correction on the multi-spectral remote sensing image collected by the first multi-spectral camera.

[0073] The quality of the UAV spectral image without radiation correction will be affected by changes in external lighting conditions. The reduction in the number and low quality of remote sensing images will have a negative impact on crop growth monitoring. In this embodiment, by obtaining the reference panel data at the same time when obtaining the multi-spectral remote sensing image of the crop planting area for single-shot real-time correction of the multi-spectral remote sensing image of the crop planting area, it has stability and real-time performance, and can effectively eliminate the influence on the multi-spectral remote sensing image under changing lighting conditions.

[0074] As another possible embodiment, input the processed set of image data into the growth analysis model for processing to obtain crop growth information, specifically including the following operations:

[0075] S401. Build a color space model based on the RGB image in the set of image data, and calculate the corresponding vegetation index of the crop planting area according to the color space model.

[0076] In this step, the gray values of the R, G, and B color channels are obtained according to the color space model corresponding to the RGB image, and the corresponding vegetation index is calculated based on the gray values of the three color channels.

[0077] S402. Extract texture features based on the multi-spectral remote sensing image in the set of image data.

[0078] Exemplarily, the texture features may include the second moment reflecting the uniformity of the image gray-scale distribution and the texture fineness, the entropy reflecting the complexity of the image gray-scale value distribution, the contrast reflecting the sharpness of the image and the texture depth, and the autocorrelation value reflecting the predictable linear relationship between the gray-scale values of two adjacent pixels, etc.

[0079] Exemplarily, the texture features can be calculated through a gray-level co-occurrence matrix.

[0080] S403. Use the obtained vegetation index and texture features as feature parameters, screen the feature parameters to obtain the final feature parameters, and construct a growth trend analysis model based on the final feature parameters.

[0081] Among them, screening the feature parameters specifically includes the following operations:

[0082] S501. Perform dimensionless processing on the feature parameters.

[0083] In this embodiment, since different feature parameters may have different units, change levels, and quantity sets, dimensionless processing is required before analysis to reduce the errors caused during the feature parameter screening process. Exemplarily, the dimensionless processing can be performed in a way of averaging, so as to have higher efficiency while maintaining the uniqueness and distribution invariance of the data.

[0084] S502. Analyze the correlation degrees among all feature parameters.

[0085] Exemplarily, the gray relational analysis method can be used to analyze the correlation degrees among all feature parameters.

[0086] S503. Construct a growth trend evaluation index system for crops, calculate the correlation degrees between each feature parameter and the growth trend evaluation index, and screen out several feature parameters with relatively weak correlation degrees among them to obtain the final feature parameters.

[0087] In this embodiment, the growth trend evaluation index system includes multiple evaluation indexes for evaluating the growth trend of crops from different dimensions. Exemplarily, the gray relational analysis method is also used in this step to calculate the correlation degrees between each feature parameter and the growth trend evaluation index. The higher the gray relational degree, the greater the influence of the feature parameter on the growth trend evaluation index. Sort each feature parameter according to the gray relational degree from high to low, and eliminate several feature parameters with relatively weak correlation degrees at the end. The remaining feature parameters are used as the final feature parameters.

[0088] S504. Calculate the geodesic distance of the mutually correlated final feature parameters through the equidistant feature mapping method, construct a geodesic distance matrix, and perform multi-dimensional scaling transformation on the geodesic distance matrix to obtain a low-dimensional feature space.

[0089] In this embodiment, the geodesic distance of the finally obtained characteristic parameters that are mutually correlated is calculated by the equidistant feature mapping method to reflect the internal mutual relationship between data. When one of the characteristic parameters changes, one or more other characteristic parameters also change in a positive or negative direction, and it can be considered that these characteristic parameters are mutually correlated. Although some characteristic parameters with a relatively low correlation degree with the growth assessment index are excluded in the above embodiment, there is still a correlation between the characteristic parameters. Therefore, it is necessary to regenerate a lower-dimensional and more independent characteristic parameter space for these characteristic parameters through a feature space transformation, so as to reduce the adverse impact of the correlation on the crop growth assessment.

[0090] As another possible embodiment, the UAV flight path is planned according to the positions of the crops with abnormal growth, which specifically includes the following operations:

[0091] S601. Calculate the representativeness of the image data corresponding to the positions of each crop with abnormal growth, where the representativeness is the ratio of the characteristic parameters when the crop has abnormal growth to the standard characteristic parameters when the crop has normal growth.

[0092] S602. Cluster the positions of each crop with abnormal growth according to the representativeness to obtain multiple clusters, and select the representative points of each cluster from them.

[0093] Exemplarily, the K-means clustering method can be used to cluster each crop with abnormal growth.

[0094] S603. Plan the UAV flight path based on all the representative points.

[0095] In this embodiment, by calculating the representativeness of the image data corresponding to each crop with abnormal growth and clustering the crops with abnormal growth based on the representativeness, the crops with abnormal growth with the same or similar representativeness are classified into the same category. When two crops with abnormal growth belong to the same cluster, it means that they have the same or similar image data characteristics, and their growth may be affected by the same factors. Then, the representative points are selected from each cluster as the image acquisition points of the UAV, and the UAV flight path is planned based on all the determined representative points, so that the UAV can collect the images of all the crops with abnormal growth that may be affected by different factors with as few points as possible, improving the data collection efficiency and reducing the requirements for the endurance performance of the UAV.

[0096] As a further possible embodiment, planning the UAV flight path based on all the representative points specifically includes the following operations:

[0097] S701. Convert the longitude and latitude coordinates of all the representative points into plane coordinates, and calculate the distance matrix between all the representative points.

[0098] S702. Take the distance matrix as the input, calculate the shortest path passing through all representative sites through a heuristic algorithm, and take the finally calculated shortest path as the UAV flight route.

[0099] Exemplarily, the heuristic algorithm can adopt a genetic algorithm, a simulated annealing algorithm, or other heuristic-type algorithms, and this embodiment does not make specific limitations on this.

[0100] This embodiment converts the longitude and latitude coordinates of the representative sites into plane coordinates, calculates the distance matrix between all representative sites, and takes the distance matrix as the input of the heuristic algorithm to calculate the shortest path passing through all representative sites, so as to quickly plan the UAV flight route with the shortest path as the goal.

[0101] As a further possible embodiment, after calculating the shortest path passing through all representative sites through the heuristic algorithm, the following operations are further included:

[0102] S801. Build a flight energy consumption model of the UAV.

[0103] The energy consumption of the UAV is usually related to factors such as flight distance, flight altitude, wind speed, and load, and the flight altitude can be converted into flight distance to a certain extent. When the UAV performs the same image acquisition task each time, its own load often does not change, so the load factor can be ignored. The energy consumption model of the UAV can be simplified as:

[0104] E = k × d

[0105] Where E represents energy consumption, k represents the energy consumption coefficient affected by the UAV performance, and d is the flight distance.

[0106] S802. With the goal of minimizing the comprehensive energy consumption, based on the real-time wind direction and wind speed and the flight energy consumption model, optimize and adjust the shortest path to obtain the final UAV flight route.

[0107] This embodiment further considers the influence of external factors such as wind direction and wind speed on the UAV energy consumption on the basis of the shortest path, optimizes and adjusts the UAV flight route, so as to further reduce the energy consumption of the UAV when performing the pest and disease image acquisition task and improve its endurance performance.

[0108] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for planning the image acquisition points of crop pests and diseases based on drones, characterized in that, The method includes: Obtaining image data of the crop planting area at different time nodes within a specified time period by using drones or satellites, and constructing a time-series-based image data set; Preprocessing the image data set, inputting the processed image data set into a growth analysis model for processing to obtain crop growth information, where the crop growth information is used to characterize the growth of crops at different positions in the crop planting area; Analyzing the crops with abnormal growth according to the crop growth information to determine the positions of the crops with abnormal growth; Planning a drone flight path according to the positions of the crops with abnormal growth, where the drone flight path includes a number of collection sites, and each collection site corresponds to a crop with abnormal growth; Controlling the drone to travel along the drone flight path, and collecting high-resolution images of the corresponding crops with abnormal growth when passing through the collection sites; Inputting the high-resolution images of the crops with abnormal growth into a pest and disease analysis model for processing to obtain pest and disease identification results, generating pest and disease prediction and prevention information and prevention suggestions according to the pest and disease identification results, and outputting them.

2. The method for planning crop pest and disease image acquisition points based on an unmanned aerial vehicle according to claim 1, wherein The image data includes RGB images and multispectral remote sensing images.

3. The method for planning the image acquisition points of crop pests and diseases based on an unmanned aerial vehicle according to claim 2, wherein, Preprocessing the image data set specifically includes the following operations: Stitching the image data in the image data set into an orthoimage; Performing geometric correction and radiometric correction on the image data in the image data set; Performing band synthesis on multispectral remote sensing images of different bands; Removing the noise in the image data and cropping the image data.

4. The method for planning the image acquisition points of crop pests and diseases based on an unmanned aerial vehicle according to claim 3, characterized in that, The radiometric correction specifically includes the following operations: Collecting multispectral remote sensing images of the crop planting area by using a first multispectral camera, and collecting downwelling light data of the same moment and the same area by using a downwelling light sensor; Collecting reference plate data of the same area by using a second multispectral camera while the first multispectral camera is taking pictures, and collecting reference plate data once every certain period of time to obtain a time-series-based reference plate data set; Performing radiometric correction on the multispectral remote sensing images collected by the first multispectral camera based on the reference plate data set and the downwelling light data.

5. A method for planning crop pest and disease image acquisition points based on an unmanned aerial vehicle according to claim 2, characterized in that, Inputting the processed image data set into a growth analysis model for processing to obtain crop growth information, specifically including the following operations: Constructing a color space model based on the RGB images in the image data set, and calculating the vegetation index of the corresponding crop planting area region according to the color space model; Extracting texture features based on the multispectral remote sensing images in the image data set; Taking the obtained vegetation index and texture features as feature parameters, screening the feature parameters to obtain final feature parameters, and constructing a growth analysis model based on the final feature parameters.

6. The method for planning crop pest and disease image acquisition points based on an unmanned aerial vehicle according to claim 5, characterized in that, Screening the feature parameters specifically includes the following operations: Performing dimensionless processing on the feature parameters; Analyzing the correlation between all feature parameters; Constructing a growth evaluation index system for crops, calculating the correlation between each feature parameter and the growth evaluation index, screening out several feature parameters with relatively weak correlation among them to obtain final feature parameters; Calculating the geodesic distance of the mutually correlated final feature parameters by using the isometric feature mapping method, constructing a geodesic distance matrix, and performing multidimensional scaling transformation on the geodesic distance matrix to obtain a low-dimensional feature space.

7. A method for planning crop pest and disease image acquisition points based on an unmanned aerial vehicle, characterized in that, Plan the UAV flight path according to the positions of the crops with abnormal growth trends, specifically including the following operations: Calculate the representativeness of the image data corresponding to the positions of each crop with abnormal growth trends. The representativeness is the ratio of the characteristic parameters when the crop growth is abnormal to the standard characteristic parameters when the crop growth is normal; Cluster the positions of each crop with abnormal growth trends according to the representativeness to obtain multiple clustering clusters, and select the representative sites of each clustering cluster; Plan the UAV flight path based on all the representative sites.

8. A method for planning crop pest and disease image acquisition points based on drones according to claim 7, characterized in that, Plan the UAV flight path based on all the representative sites, specifically including the following operations: Convert the longitude and latitude coordinates of all the representative sites into plane coordinates, and calculate the distance matrix between all the representative sites; Take the distance matrix as the input, calculate the shortest path passing through all the representative sites through a heuristic algorithm, and take the finally calculated shortest path as the UAV flight path.

9. The method for planning crop pest and disease image acquisition points based on an unmanned aerial vehicle according to claim 8, characterized in that, After calculating the shortest path passing through all the representative sites through the heuristic algorithm, the following operations are also included: Construct a flight energy consumption model for the UAV; With the goal of minimizing the comprehensive energy consumption, optimize and adjust the shortest path based on the real-time wind direction and speed and the flight energy consumption model to obtain the final UAV flight path.

Images