Ecological restoration engineering supervision method and device based on unmanned aerial vehicle, and medium

Abstract

The invention discloses an ecological restoration engineering supervision method and device based on an unmanned aerial vehicle and a medium, and the method comprises the steps: controlling the unmanned aerial vehicle and a satellite to collect image data of a project supervision region at the same time, and obtaining unmanned aerial vehicle data and satellite remote sensing data; wherein the take-off and landing point of the unmanned aerial vehicle is determined by identifying the ecological change trend of a project supervision area; correcting the satellite remote sensing data based on the unmanned aerial vehicle data to obtain multispectral reflectivity data; and evaluating the ecological restoration effect of the project supervision area according to the multispectral reflectivity data to obtain an ecological restoration project supervision result. According to the ecological restoration engineering supervision method and device based on the unmanned aerial vehicle and the medium, the unmanned aerial vehicle and the satellite are controlled to collect the remote sensing data of the project supervision area at the same time, scientific processing and analysis are carried out based on the data, the comprehensiveness of the data is guaranteed, the data collection and analysis efficiency is improved, and the method is suitable for popularization and application. The problem that it is difficult to quickly and accurately supervise the ecological restoration project can be solved.

Description

Technical Field

[0001] The present invention relates to the technical field of environmental monitoring, and in particular to a method, device and medium for monitoring an ecological restoration project based on an unmanned aerial vehicle. Background Art

[0002] Landscape and ecological restoration projects cover a wide range of areas. The project forms in different regions are complex and varied. The supervision cycle is long and requires high-precision and high-temporal resolution data support. Landscape and ecological restoration projects have significantly improved the regional environmental quality, effectively controlling dust and harmful gases in the air, purifying water quality, reducing noise pollution, and greatly improving the living environment of residents. Multispectral remote sensing technology has a wide range of applications in ecology. Multispectral image data obtained by satellites can be used to monitor environmental pollution sources such as the atmosphere, water bodies and soil. By analyzing the concentration and distribution of pollutants in image data, the degree and scope of environmental pollution can be evaluated, providing a basis for environmental protection and governance.

[0003] However, multispectral remote sensing technology is limited by factors such as resolution, cloud obstruction, light, temperature, soil and atmosphere, which can easily lead to data fluctuations. Even if ground measurement and correction are carried out later, the measurement data can easily become invalid because the same measurement conditions cannot be restored. In addition, it takes a long time to supervise ecological restoration projects under conditions such as vegetation status, temperature, light, soil moisture and atmosphere, which can easily lead to correction failures or low data credibility. Summary of the invention

[0004] The present invention provides an ecological restoration project supervision method, device and medium based on a drone to solve the problem that it is difficult to quickly and accurately supervise the ecological restoration project.

[0005] Controlling the drone and the satellite to simultaneously collect image data of the project supervision area, and obtaining drone data and satellite remote sensing data; wherein the take-off and landing points of the drone are determined by identifying the ecological change trend and potential problem areas of the project supervision area;

[0006] Correcting the satellite remote sensing data based on the drone data to obtain multispectral reflectance data;

[0007] The ecological restoration effect of the project supervision area is evaluated based on the multispectral reflectance data to obtain the ecological restoration project supervision results.

[0008] In the present invention, the high-resolution image data of the UAV can clearly reflect the surface details and help to find potential problem areas; while the satellite remote sensing data can reveal the changing trend of surface features through time series analysis; by combining the data of the two, they can verify and correct each other to improve the accuracy and reliability of the data. The UAV has the ability of rapid deployment and real-time monitoring, and can monitor key areas multiple times in a short period of time, thereby improving the monitoring efficiency; and although the acquisition cycle of satellite remote sensing data is relatively long, once acquired, it can cover a large area and reduce duplication of work; the combination of the two makes the monitoring work more efficient. In addition, by identifying ecological change trends and potential problem areas, the areas that the UAV needs to focus on can be accurately determined, so that take-off and landing points can be set in these areas. This precise positioning reduces aimless and large-scale flights, so the data collection capabilities of the UAV can be maximized to ensure that key areas are quickly monitored and evaluated.

[0009] Compared with the existing technology, the present invention controls drones and satellites to simultaneously collect remote sensing data of the project supervision area, and conducts scientific processing and analysis based on these data, which not only ensures the comprehensiveness of the data, but also improves the efficiency of data collection and analysis. Therefore, it can solve the problem of difficulty in quickly and accurately supervising ecological restoration projects.

[0010] As a preferred solution, the take-off and landing points of the drone are determined by identifying the ecological change trends and potential problem areas in the project supervision area, specifically:

[0011] Dividing the project supervision area into a two-dimensional grid;

[0012] By performing curve compression and clustering processing on the two-dimensional grid, the coefficient of variation of the subclass grids in the two-dimensional grid is calculated;

[0013] The subclass grids in the two-dimensional grid whose coefficient of variation exceeds a preset threshold are defined as abnormal regions, to obtain an abnormal region set;

[0014] The take-off and landing points of the UAV are obtained by performing spatial superposition processing on the abnormal areas in the abnormal area concentration.

[0015] This preferred solution can accurately locate abnormal areas by calculating the coefficient of variation of subclass grids, providing an accurate basis for the subsequent selection of drone take-off and landing points. Curve compression and clustering processing algorithms can quickly process large amounts of data and quickly identify abnormal areas where the coefficient of variation exceeds the preset threshold.

[0016] As a preferred solution, by performing curve compression and clustering processing on the two-dimensional grid, the coefficient of variation of the subclass grids in the two-dimensional grid is calculated, specifically:

[0017] Obtaining the normalized vegetation index of the subclass grid from the two-dimensional grid to obtain initial time series data;

[0018] Performing curve compression on the initial time series data to obtain first time series data;

[0019] Based on the first set of time series data points in the first time series data, connecting the first point and the last point into a straight line to obtain a target straight line;

[0020] Calculating the distances from all points in the first set of time series data points to the target straight line to obtain a distance set;

[0021] Based on the maximum value in the distance set, target data of the time series data is eliminated to obtain thinned second time series data;

[0022] The coefficient of variation of the subclass grid in the two-dimensional grid is calculated based on the second time series data.

[0023] This preferred solution can obtain more compact first time series data by curve compression of the initial time series data, reduce the amount of data, and improve the efficiency of subsequent data processing. The thinning process based on the distance set and the target line can eliminate redundant data points and retain key information, thereby reducing the amount of calculation while ensuring data accuracy.

[0024] As a preferred solution, based on the maximum value in the distance set, the target data of the time series data is eliminated to obtain the thinned second time series data, specifically:

[0025] Obtaining the maximum value in the distance set to obtain a maximum distance value;

[0026] If the maximum distance value is less than a preset threshold, other points except the first and last points in the time series curve are removed; wherein the time series curve is composed of the time series data;

[0027] If the maximum distance value is greater than or equal to the preset threshold, retaining the first data point corresponding to the maximum distance value, and dividing the first group of time series data points into two sub-point sets with the first data point as the boundary;

[0028] The two sub-point sets are processed recursively, and the thinned second time series data is obtained by traversing a plurality of time series data point sets in the time series data.

[0029] This preferred solution can significantly reduce the amount of data and improve the efficiency of subsequent processing and analysis by eliminating points in the time series curve that contribute less to the curve shape. By recursively processing the segmented sub-point set, it can ensure that the reconstructed time series curve can better maintain the shape and characteristics of the original curve.

[0030] As a preferred solution, the coefficient of variation of the subclass grid in the two-dimensional grid is calculated according to the second time series data, specifically:

[0031] Controlling the subclass grids in the second time series data to cluster with the subclass grids within a preset radius to obtain a clustering result;

[0032] Based on the values ​​of the preset time points in the clustering results, spatially clustering the data of several categories in the clustering results to obtain a comprehensive subclass grid set;

[0033] The coefficient of variation of the comprehensive subclass grid set in the two-dimensional grid is obtained by calculating the degree of discreteness of the data distribution of the comprehensive subclass grid set.

[0034] This preferred solution clusters similar time series subclass grids into categories through the clustering of time series data, which helps to identify areas with similar dynamic change characteristics; spatial clustering is performed based on the values ​​at specific time points in the clustering results, which can further integrate similar areas in space to form a comprehensive subclass grid set; through two clusterings, the scope of clustering can be gradually narrowed and the accuracy of clustering can be improved. The coefficient of variation is obtained by calculating the degree of data distribution dispersion of the comprehensive subclass grid set. This indicator can intuitively reflect the degree of difference between the subclass grids in the comprehensive subclass grid set, providing strong support for subsequent analysis and decision-making.

[0035] As a preferred solution, the take-off and landing points of the UAV are obtained by spatially superimposing the abnormal areas in the abnormal area concentration, specifically:

[0036] According to the value of the coefficient of variation, the abnormal region set is divided into a number of abnormal region categories;

[0037] Randomly select a number of abnormal regions from the several abnormal region categories to obtain data samples;

[0038] Simplifying the subclass grids in the data sample into a uniform dot matrix to obtain a dot matrix set;

[0039] Taking each point in the dot matrix as the center of a circle, a plurality of circular buffer zones are established according to a preset radius;

[0040] The plurality of circular buffer zones are spatially superimposed to obtain a numerical distribution of preset objective function values ​​corresponding to each zone;

[0041] Based on the numerical distribution, several areas with the highest values ​​are selected as the take-off and landing points of the UAV.

[0042] This preferred solution divides the abnormal area set according to the numerical value of the coefficient of variation, and can objectively and accurately identify areas with different abnormal degrees and classify them; randomly extracting a number of abnormal areas from the divided abnormal area categories as data samples not only ensures the representativeness of the samples, but also avoids the subjectivity of human selection. The subclass grid in the data sample is simplified into a uniform dot matrix, which simplifies the complexity of the data and improves processing efficiency. Taking each point in the dot matrix set as the center of the circle and establishing a circular buffer according to a preset radius can intuitively reflect the spatial characteristics around each point for spatial superposition and analysis.

[0043] As a preferred solution, the drone and satellite are controlled to simultaneously collect image data of the project supervision area to obtain drone data and satellite remote sensing data, specifically:

[0044] Adjusting the observation angle and camera parameters of the drone to be consistent with the observation conditions of the data acquired by the satellite;

[0045] Controlling the UAV and the satellite to simultaneously collect image data of the project supervision area to obtain the UAV data and the satellite remote sensing data;

[0046] If there are areas where data is missing or where the emissivity changes exceed a preset value in the satellite remote sensing data, supplementary photography or encrypted measurement is performed by the drone.

[0047] In this preferred solution, satellite remote sensing data has the advantages of macroscopicity, continuity and timeliness, but may be missing or have reduced quality due to factors such as cloud cover and ground cover; drones can flexibly adjust flight routes and altitudes to conduct supplementary photography or encrypted measurements of specific areas, thereby filling gaps in satellite data or improving data accuracy.

[0048] As a preferred solution, the project supervision area is specifically:

[0049] Obtain basic geographic information of the comprehensive project supervision area;

[0050] Based on the basic geographic information, by identifying the distribution of different land types and soil types, the key monitoring areas in the comprehensive project supervision area are determined to obtain the project supervision area.

[0051] This preferred solution determines the key monitoring areas by identifying the distribution of different land types and soil types, and can delineate the required regulatory areas to avoid the problem of wasting regulatory resources.

[0052] The present application also provides a drone-based ecological restoration project supervision device, including a data module, a correction module, and an evaluation module;

[0053] The data module is used to control the drone and the satellite to simultaneously collect image data of the project supervision area, thereby obtaining drone data and satellite remote sensing data; wherein the take-off and landing points of the drone are determined by identifying the ecological change trends and potential problem areas of the project supervision area;

[0054] The correction module is used to correct the satellite remote sensing data based on the UAV data to obtain multispectral reflectance data;

[0055] The evaluation module is used to evaluate the ecological restoration effect of the project supervision area according to the multispectral reflectance data to obtain the ecological restoration project supervision results.

[0056] The present application also provides a storage medium having a computer program stored thereon, and the computer program is called and executed by a computer to implement the above-mentioned method for supervising ecological restoration projects based on drones. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 It is a flow chart of a method for monitoring an ecological restoration project based on a drone provided in an embodiment of the present application;

[0058] Figure 2 is a schematic diagram of spatial superposition provided in an embodiment of the present application;

[0059] Figure 3 It is a schematic diagram of space buffering provided in an embodiment of the present application;

[0060] Figure 4 This is a schematic diagram of the result of superposition value accumulation provided in the embodiment of the present application;

[0061] Figure 5 It is a structural schematic diagram of an ecological restoration project monitoring device based on a drone provided in an embodiment of the present application. DETAILED DESCRIPTION

[0062] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0063] In the description of this application, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of this application, unless otherwise specified, "several" means two or more.

[0064] The embodiment of the present application provides a drone-based ecological restoration project supervision method, which is mainly used in situations where it is necessary to balance data accuracy, timeliness and cost to achieve full-cycle efficient supervision of landscape and ecological restoration projects.

[0065] Embodiment 1:

[0066] See also Figure 1 The embodiment of the present application provides a method for monitoring ecological restoration projects based on drones, including S1 to S3, and the specific implementation steps are as follows:

[0067] S1. Control the UAV and satellite to simultaneously collect image data of the project supervision area to obtain UAV data and satellite remote sensing data; wherein, the take-off and landing points of the UAV are determined by identifying the ecological change trends and potential problem areas in the project supervision area.

[0068] Step S1 of the embodiment of the present application includes S1.1 to S1.7; wherein S1.1 is a process of obtaining a project supervision area, S1.2 is a process of establishing a distance set based on the project supervision area, S1.3 is a process of obtaining second time series data based on the distance set, S1.4 is a process of calculating a coefficient of variation based on the second time series data, S1.5 is a process of selecting the best airport placement candidate area based on the coefficient of variation, S1.6 is a process of determining a drone take-off and landing point in combination with the best airport placement candidate area and road network data, and S1.7 is a process of controlling drones and satellites to collect data, specifically:

[0069] S1.1. Obtain basic geographic information of the comprehensive project supervision area, including topography, landforms, water systems and vegetation coverage;

[0070] Collect land planning and soil type data provided by relevant departments, and establish a unified database management system to store and manage all data by integrating records and performance data of historical ecological restoration projects;

[0071] The comprehensive data, including the basic geographic information of the comprehensive project supervision area and the data in the database management system, are imported into the GIS platform (Geographic Information System). The comprehensive data are spatially overlaid and analyzed using GIS technology. The project supervision area is obtained by identifying the distribution of different land types and soil types, dividing the supervision area according to NDVI data (Normalized Difference Vegetation Index) and ecological sensitivity, and determining the key monitoring areas.

[0072] In this embodiment S1.1, by identifying the distribution of different land types and soil types, the key monitoring areas can be determined, and the required regulatory areas can be delineated to avoid the problem of wasting regulatory resources.

[0073] S1.2. Divide the project supervision area into two-dimensional grids; the size of each subclass grid in the two-dimensional grid corresponds to a pixel of the multispectral satellite, that is, a subclass grid is as large as a pixel of NDVI; and "Prow,col[n]" represents the NDVI value of the grid corresponding to row th row and col th column at the nth measurement, and the sampling interval is T;

[0074] Obtain the NDVI values ​​of the subclass grids from the two-dimensional grid to obtain the initial time series data;

[0075] Using a curve compression algorithm (such as Douglas-Peucker algorithm and Visvalingam-Whyatt algorithm) to perform curve compression on the initial time series data to obtain the first time series data;

[0076] Based on the first set of time series data points P in the first time series data, the first point p0 and the last point p n Connect them into a straight line to get the target straight line AB; where P = {p0, p1, ..., p n}, and for each point: p i =(x i ,y i ), for the target line AB, we have: A(x0,y0), B(x n ,y n );

[0077] Calculate all points p in the first set of time series data points i (i=1,2,...,n-1) The distance d to the target line AB i , and get the distance set.

[0078] Among them, the distance formula from a point to the target line is:

[0079]

[0080] The general equation of the target straight line is:

[0081] Ax + By + C = 0

[0082] For the parameters A, B, and C, there are:

[0083] A = y n - y0;

[0084] B = x0 - x n ;

[0085] C = x n y0 - x0y n

[0086] Wherein, x and y respectively represent the abscissa and ordinate of a certain point on the target straight line.

[0087] In this embodiment S1.2, due to the high time resolution, large amount of data and data redundancy, direct clustering will consume too much time. Therefore, curve compression is performed before clustering, which can improve the clustering efficiency on the premise of retaining the shape of the time curve.

[0088] S1.3. Obtain the maximum value in the distance set to get the maximum distance value d max ;

[0089] If the maximum distance value d max is less than the preset threshold D, that is, d max < D, then remove other points in the time series curve except the first point p0 and the last point p n ; Among them, the time series curve is composed of time series data;

[0090] If the maximum distance value d max is greater than or equal to the preset threshold D, that is, d max ≥ D, then retain the first data point p max corresponding to the maximum distance value d max_index , and divide the first set of time series data point set P into two sub-point sets with the first data point p max_index as the boundary; Among them, the two sub-point sets are P1 = {p0,..., p max_index} and P2 = {p max_index ,..., p n};

[0091] Repeat the above curve clustering and data removal calculation operations on the two sub-point sets for recursive processing; by traversing several groups of time series data point sets in the time series data, after the recursive processing ends, merge all the retained points in order to obtain the thinned second time series data.

[0092] In this embodiment S1.3, by removing points in the time series curve that contribute less to the curve shape, the amount of data can be significantly reduced, thereby improving the efficiency of subsequent processing and analysis. By recursively processing the segmented sub-point set, it can be ensured that the reconstructed time series curve can better maintain the shape and characteristics of the original curve;

[0093] Therefore, the thinning process based on the distance set and the target line can eliminate redundant data points and retain key information, thereby reducing the amount of calculation while ensuring data accuracy.

[0094] S1.4, controlling the subclass grids in the second time series data to cluster with the subclass grids within a preset radius R1 to obtain a clustering result;

[0095] Based on the value of the last time point in the clustering result, spatial clustering is performed on the data of several categories in the clustering result to obtain a comprehensive subclass grid set consisting of many spatially clustered small classes under each large class;

[0096] By calculating the discrete degree of data distribution of the comprehensive subclass grid set, the coefficient of variation CV of the comprehensive subclass grid set in the two-dimensional grid is obtained.

[0097] The calculation formula of the coefficient of variation is:

[0098] CV=(SD / Mean)×100%

[0099] Among them, SD is the standard deviation of the NDVI value of each subclass grid, which is used to reflect the fluctuation of the NDVI value within the subclass grid; Mean is the average value of all NDVI values ​​in the comprehensive subclass grid set, which is used to reflect the average vegetation coverage of the subclass grid at a specific time point.

[0100] In this embodiment S1.4, similar time series subclass grids are clustered into categories through the clustering of time series data, which helps to identify areas with similar dynamic change characteristics; spatial clustering is performed based on the values ​​at specific time points in the clustering results, which can further integrate similar areas in space to form a comprehensive subclass grid set; through two clusterings, the scope of clustering can be gradually narrowed and the accuracy of clustering can be improved. The coefficient of variation is obtained by calculating the degree of data distribution dispersion of the comprehensive subclass grid set. This indicator can intuitively reflect the degree of difference between the subclass grids in the comprehensive subclass grid set, providing strong support for subsequent analysis and decision-making.

[0101] S1.5, defining the subclass grids whose coefficient of variation CV in the two-dimensional grid exceeds a preset threshold as abnormal regions, and obtaining an abnormal region set;

[0102] According to the value of the coefficient of variation CV and the importance of the region, the abnormal region set is divided into several abnormal region categories; wherein the several abnormal region categories include three categories: strong, medium and weak;

[0103] A number of abnormal regions are randomly selected from a number of abnormal region categories to obtain data samples S = {s1, s2, s3}; wherein a sample point in the data sample S corresponds to a sample point in a subclass grid in a number of abnormal region categories; and the sample point has a certain flexibility in the subclass grid, that is, it can move within the grid;

[0104] The minimum bounding circle algorithm is used to plan the flight path of the drone to ensure that the drone can cover as many different types of sample points as possible; the drone can collect samples in a shorter time under the condition of a certain number of samples, ensuring the synchronization of data;

[0105] Simplify the subclass grids in the data sample S into a uniform lattice to obtain a lattice set;

[0106] With each point in the dot matrix as the center, a number of circular buffers are established according to a preset radius R2; wherein the value of each buffer in the number of circular buffers is the CV value (coefficient of variation) of the sample point;

[0107] Several circular buffers are spatially superimposed to divide the buffers from each other. All values ​​are added together while controlling that the same subclass ID (Identifier) ​​is added only once to obtain the numerical distribution of the preset objective function value F(C) corresponding to each area.

[0108] Based on the numerical distribution, the N areas with the highest values ​​are selected as the best airport placement candidate areas for drones.

[0109] Among them, the objective function value is:

[0110]

[0111] Among them, α i is the weight of the importance of the i-th subclass of the sample, β i is the CV value of the sample point, K is the number of all subclasses; i A Boolean variable that is 1 if the subclass is overridden and 0 otherwise.

[0112] To apply the embodiments of this application, please refer to Figure 2-4 ;

[0113] Figure 2is a schematic diagram of spatial superposition provided in an embodiment of the present application, which shows the principle of spatial superposition of a plurality of circular buffer zones, namely: when two graphics are superimposed, the intersection becomes an independent graphic;

[0114] Figure 3 is a schematic diagram of spatial buffering provided in an embodiment of the present application, which shows a process of spatially superimposing a plurality of circular buffer zones;

[0115] Figure 4 It is a schematic diagram of the result of superposition value accumulation provided in an embodiment of the present application, which shows the result of superposition value accumulation obtained after spatial superposition of several circular buffer zones.

[0116] In this embodiment S1.5, the abnormal region set is divided according to the numerical value of the coefficient of variation, and regions of different abnormality degrees can be objectively and accurately identified and classified; a number of abnormal regions are randomly selected from the divided abnormal region categories as data samples, which not only ensures the representativeness of the samples, but also avoids the subjectivity of human selection. The subclass grid in the data sample is simplified into a uniform dot matrix, which simplifies the complexity of the data and improves the processing efficiency. A circular buffer zone is established with each point in the dot matrix as the center and according to a preset radius, which can intuitively reflect the spatial characteristics around each point for spatial superposition and analysis.

[0117] S1.6. Obtain detailed road network data within a preset area, including information such as road geometry, type and traffic restrictions;

[0118] Overlay the location of the best airport placement candidate area with the road network data to determine the optimal drone take-off point;

[0119] Use an efficient path search algorithm (such as Dijkstra algorithm) combined with the UAV flight path planned using the minimum circle cover algorithm to calculate the shortest path from the selected optimal UAV take-off point to all destination sample points;

[0120] Based on the calculation results of the shortest path and the actual operation requirements of the UAV, the final layout point of the UAV airport is determined by comprehensively considering factors such as the location of the take-off point, the surrounding environment and traffic conditions, thereby obtaining the take-off and landing points of the UAV;

[0121] Fixed drone airports and mobile drone airports are set up based on the take-off and landing points of drones; and ensure that the facilities of the drone airports are complete and can meet the needs of drone take-off and landing and maintenance; among them, fixed drone airports are set up in preset key supervision areas, and mobile drone airports are set up according to actual needs.

[0122] S1.7. Before the satellite passes overhead, calibrate the satellite's multispectral camera using radiation correction plates with different reflectivity of 25%, 50% and 75% to ensure its stable performance and high accuracy; and record the calibration results for subsequent data processing and analysis; where the satellite is a multispectral satellite;

[0123] Control the drone to depart from the drone airport (take-off and landing point) and arrive at the destination sample point; adjust the drone's observation angle (azimuth) and camera parameters (camera pitch angle) to be consistent with the observation conditions of the data obtained by the satellite;

[0124] Control the UAV and satellite to simultaneously collect image data of the project supervision area, and obtain UAV data and high-resolution satellite remote sensing data; among which, the image data is multispectral image data;

[0125] If there are areas with missing data in the satellite remote sensing data or areas where the emissivity changes exceed the preset value, supplementary photography or encrypted measurements will be carried out through drones to ensure the integrity and accuracy of the data.

[0126] In this embodiment S1.7, satellite remote sensing data has the advantages of macroscopicity, continuity and timeliness, but may be missing or have reduced quality due to factors such as cloud cover and ground cover; drones can flexibly adjust flight routes and altitudes, conduct supplementary photography or encrypted measurements of specific areas, thereby filling gaps in satellite data or improving data accuracy.

[0127] S2. Correct satellite remote sensing data based on UAV data to obtain multispectral reflectance data.

[0128] Step S2 of the embodiment of the present application is specifically as follows:

[0129] Perform preliminary preprocessing on drone data, such as denoising and enhancement, and extract reflectivity information from drone data for subsequent data analysis;

[0130] Time-synchronize drone data with satellite remote sensing data to ensure data consistency and comparability;

[0131] Based on the actual radiation reflectivity of each band in the UAV data, the satellite remote sensing data is corrected to eliminate the differences caused by atmospheric correction and obtain multispectral reflectivity data; the accuracy and reliability of the multispectral reflectivity data are evaluated;

[0132] The relevant index (such as NDVI) is calculated based on the multispectral reflectance data, and the latest relevant index is used as the basis for the next sampling.

[0133] S3. Evaluate the ecological restoration effect of the project supervision area based on multispectral reflectance data and obtain the ecological restoration project supervision results.

[0134] Step S3 of the embodiment of the present application includes S3.1 to S3.2; wherein S3.1 is the process of obtaining the supervision results of the ecological restoration project, and S3.2 is the process of continuous monitoring and feedback, specifically:

[0135] S3.1. By conducting time series analysis and spatial distribution analysis on the NDVI data after time series fusion in the multispectral reflectance data, the ecological restoration effect of the project supervision area is evaluated and the first supervision result is obtained;

[0136] The first supervision result is combined with the field investigation results of the project supervision area and the preset expert knowledge to identify the existing problems and challenges and obtain the supervision results of the ecological restoration project.

[0137] S3.2. Develop a regular monitoring plan to ensure continuous monitoring of the project supervision area; conduct irregular spot checks on the project supervision area based on actual needs and monitoring results; dynamically adjust the sampling plan and supervision focus based on monitoring results and the progress of the ecological restoration project;

[0138] Establish a real-time data transmission system to transmit monitoring data to the supervision system in real time; use data visualization technology to convert monitoring data into intuitive charts and images; and display the progress and results of ecological restoration projects in real time through visualization;

[0139] Establish an early warning system, set early warning thresholds and trigger conditions; conduct real-time monitoring and analysis of monitoring data to identify abnormal situations; once an abnormal situation is found, issue an early warning signal in a timely manner and respond;

[0140] Collect and analyze monitoring data to evaluate the progress and effectiveness of ecological restoration projects; regularly prepare monitoring reports to summarize monitoring results and problems found; submit monitoring reports to project management departments and relevant policy-making agencies to provide a scientific basis for project management and policy making.

[0141] Overall, this embodiment has the following beneficial effects:

[0142] In the present invention, the high-resolution multispectral image data of the UAV can clearly reflect the surface details and help to find potential problem areas; and the satellite remote sensing data can reveal the changing trends of surface features through time series analysis; by combining the data of the two, they can verify and correct each other to improve the accuracy and reliability of the data. The UAV has the ability to be quickly deployed and monitored in real time, and can monitor key areas multiple times in a short period of time, thereby improving monitoring efficiency; and although the acquisition cycle of satellite remote sensing data is relatively long, once acquired, it can cover a large area and reduce duplication of work; the combination of the two makes monitoring more efficient. In addition, by identifying ecological change trends and potential problem areas, it is possible to accurately determine the areas where the UAV needs to focus, so that take-off and landing points can be set in these areas. This precise positioning reduces aimless and large-scale flights, so the data collection capabilities of the UAV can be maximized to ensure that key areas are quickly monitored and evaluated;

[0143] Through the deep integration and correction technology of drone and satellite remote sensing data, full-cycle, high-precision, and low-cost supervision of landscape and ecological restoration projects has been achieved. This method not only significantly improves the accuracy and timeliness of data, but also effectively reduces the cost of human verification and enhances the intelligence and automation level of supervision. At the same time, through continuous monitoring and feedback mechanisms, problems arising in the ecological restoration process can be discovered and resolved in a timely manner to ensure the smooth implementation of the project and achieve the expected results.

[0144] Embodiment 2:

[0145] See also Figure 5 , the embodiment of the present application provides an ecological restoration project supervision device based on a drone, including a data module 10, a correction module 20 and an evaluation module 30;

[0146] The data module 10 is used to control the drone and the satellite to simultaneously collect image data of the project supervision area, thereby obtaining drone data and satellite remote sensing data; wherein the take-off and landing points of the drone are determined by identifying the ecological change trends and potential problem areas in the project supervision area;

[0147] A correction module 20, for correcting satellite remote sensing data based on UAV data to obtain multispectral reflectance data;

[0148] The evaluation module 30 is used to evaluate the ecological restoration effect of the project supervision area based on the multi-spectral reflectance data to obtain the ecological restoration project supervision results.

[0149] In one embodiment, the data module 10 includes an area unit, a distance unit, a sequence unit, a variation unit, a candidate unit, a take-off and landing unit, and a collection unit; wherein the area unit is a process of obtaining a project supervision area, the distance unit is a process of establishing a distance set based on the project supervision area, the sequence unit is a process of obtaining the second time series data according to the distance set, the variation unit is a process of calculating the coefficient of variation based on the second time series data, the candidate unit is a process of selecting the best airport placement candidate area based on the coefficient of variation, the take-off and landing unit is a process of determining the take-off and landing point of the drone in combination with the best airport placement candidate area and road network data, and the collection unit is a process of controlling the drone and the satellite to collect data, specifically:

[0150] Regional unit, used to obtain basic geographic information of the comprehensive project supervision area, including topography, landforms, water systems and vegetation cover;

[0151] The regional unit is also used to collect land planning and soil type data provided by relevant departments, and to establish a unified database management system to store and manage all data by integrating the records and effectiveness data of historical ecological restoration projects;

[0152] The regional unit is also used to import comprehensive data, including basic geographic information of the comprehensive project supervision area and data in the database management system, into the GIS platform (Geographic Information System); use GIS technology to perform spatial overlay analysis on the comprehensive data, and obtain the project supervision area by identifying the distribution of different land types and soil types, dividing the supervision area according to NDVI data (Normalized Difference Vegetation Index) and ecological sensitivity and determining the key monitoring areas.

[0153] The regional unit of this embodiment determines the key monitoring area by identifying the distribution of different land types and soil types, and can delineate the required supervision area to avoid the problem of wasting supervision resources.

[0154] Distance unit, used to divide the project supervision area into two-dimensional grids; the size of each subclass grid in the two-dimensional grid corresponds to a pixel of the multispectral satellite, that is, a subclass grid is as large as a pixel of NDVI; and "Prow,col[n]" represents the NDVI value of the grid corresponding to row row and col column at the nth measurement, and the sampling interval is T;

[0155] The distance unit is also used to obtain the NDVI values ​​of the subclass grids from the two-dimensional grid to obtain the initial time series data;

[0156] The distance unit is also used to use a curve compression algorithm (such as Douglas-Peucker algorithm and Visvalingam-Whyatt algorithm) to perform curve compression on the initial time series data to obtain the first time series data;

[0157] The distance unit is also used to separate the first point p0 and the last point p based on the first set of time series data points P in the first time series data. n Connect them into a straight line to get the target straight line AB; where P = {p0, p1, ..., p n}, and for each point: p i =(x i ,y i ), for the target line AB, we have: A(x0,y0), B(x n ,y n );

[0158] The distance unit is also used to calculate all points p in the first set of time series data points. i (i=1,2,...,n-1) The distance d to the target line AB i , and get the distance set.

[0159] Among them, the distance formula from a point to the target line is:

[0160]

[0161] The general equation of the target line is:

[0162] Ax+By+C=0

[0163] For parameters A, B and C we have:

[0164] A=y n -y0;

[0165] B=x0-x n ;

[0166] C=x n y0-x0y n

[0167] Among them, x and y represent the horizontal and vertical coordinates of a point on the target line respectively.

[0168] In the distance unit of this embodiment, due to the high time resolution, large data volume and data redundancy, direct clustering will take too much time, so curve compression is performed before clustering, which can improve clustering efficiency while retaining the shape of the time curve.

[0169] Sequence unit, used to obtain the maximum value in the distance set and obtain the maximum distance value d max ;

[0170] The sequence unit is also used to, if the maximum distance value d max is less than a preset threshold D, i.e., d max < D, then remove all points in the time series curve except the first point p0 and the last point p n ; among them, the time series curve is composed of time series data;

[0171] The sequence unit is also used to, if the maximum distance value d max is greater than or equal to the preset threshold D, i.e., d max ≥ D, then retain the first data point p max corresponding to the maximum distance value d max_index , and divide the first set of time series data point sets P into two sub-point sets with the first data point p max_index as the boundary; among them, the two sub-point sets are P1 = {p0,..., p max_index} and P2 = {p max_index ,..., p n};

[0172] The sequence unit is also used to repeatedly perform the above curve clustering and data removal calculation operations on the two sub-point sets for recursive processing; by traversing several groups of time series data point sets in the time series data, after the recursive processing ends, all the retained points are merged in order to obtain the thinned second time series data.

[0173] In this embodiment, the sequence unit can significantly reduce the data volume and improve the efficiency of subsequent processing and analysis by removing the points with less contribution to the curve shape in the time series curve. By recursively processing the divided sub-point sets, it can ensure that the reconstructed time series curve can better maintain the shape and characteristics of the original curve;

[0174] Therefore, the thinning process based on the distance set and the target line can remove redundant data points and retain key information, thereby reducing the amount of calculation while ensuring data accuracy.

[0175] The mutation unit is used to control the clustering of the sub-class grids in the second time series data with the sub-class grids within a preset radius R1 to obtain a clustering result;

[0176] The mutation unit is also used to perform spatial clustering on the data of several categories in the clustering result based on the value of the last time point in the clustering result to obtain a comprehensive sub-class grid set composed of many spatially aggregated small classes under each large class;

[0177] The mutation unit is also used to calculate the data distribution dispersion degree of the comprehensive sub-class grid set to obtain the coefficient of variation CV of the comprehensive sub-class grid set in the two-dimensional grid.

[0178] The calculation formula of the coefficient of variation is:

[0179] CV=(SD / Mean)×100%

[0180] Among them, SD is the standard deviation of the NDVI value of each subclass grid, which is used to reflect the fluctuation of the NDVI value within the subclass grid; Mean is the average value of all NDVI values ​​in the comprehensive subclass grid set, which is used to reflect the average vegetation coverage of the subclass grid at a specific time point.

[0181] The variation unit of this embodiment clusters similar time series subclass grids into categories through clustering of time series data, which helps to identify areas with similar dynamic change characteristics; spatial clustering is performed based on the values ​​at specific time points in the clustering results, which can further integrate similar areas in space to form a comprehensive subclass grid set; through two clusterings, the scope of clustering can be gradually narrowed and the accuracy of clustering can be improved. The coefficient of variation is obtained by calculating the degree of data distribution dispersion of the comprehensive subclass grid set. This indicator can intuitively reflect the degree of difference between the subclass grids in the comprehensive subclass grid set, providing strong support for subsequent analysis and decision-making.

[0182] The candidate unit is used to define the subclass grids whose coefficient of variation CV in the two-dimensional grid exceeds a preset threshold as abnormal regions, and obtain an abnormal region set;

[0183] The candidate unit is also used to divide the abnormal region set into a number of abnormal region categories according to the numerical value of the coefficient of variation CV and the importance of the region; wherein the categories of the several abnormal region categories include three categories: strong, medium and weak;

[0184] The candidate unit is also used to randomly extract several abnormal regions from several abnormal region categories to obtain data samples S = {s1, s2, s3}; wherein a sample point in the data sample S corresponds to a sample point in a subclass grid in several abnormal region categories; and the sample point has a certain flexibility in the subclass grid, that is, it can move in the grid;

[0185] Candidate units are also used to plan the flight path of the drone using the Minimum Bounding Circle Algorithm to ensure that the drone can cover as many different types of sample points as possible; so that the drone can collect samples in a shorter time under the condition of a certain number of samples, ensuring the synchronization of data;

[0186] The candidate unit is also used to simplify the subclass grid in the data sample S into a uniform point matrix to obtain a point matrix set;

[0187] The candidate unit is further used to establish a plurality of circular buffers with each point in the dot matrix as the center of the circle according to a preset radius R2; wherein the value of each buffer in the plurality of circular buffers is the CV value (coefficient of variation) of the sample point;

[0188] The candidate unit is also used to spatially superimpose several circular buffers so that the buffers are mutually divided, and all values ​​are added under the condition that the same subclass ID (Identifier) ​​is only added once to obtain the numerical distribution of the preset objective function value F(C) corresponding to each area;

[0189] The candidate unit is also used to select N areas with the highest values ​​as the best airport placement candidate areas for drones based on the numerical distribution.

[0190] Among them, the objective function value is:

[0191]

[0192] Among them, α i is the weight of the importance of the i-th subclass of the sample, β i is the CV value of the sample point, K is the number of all subclasses; i A Boolean variable that is 1 if the subclass is overridden and 0 otherwise.

[0193] To apply the embodiments of this application, please refer to Figure 2-4 ;

[0194] Figure 2 is a schematic diagram of spatial superposition provided in an embodiment of the present application, which shows the principle of spatial superposition of a plurality of circular buffer zones, namely: when two graphics are superimposed, the intersection becomes an independent graphic;

[0195] Figure 3 is a schematic diagram of spatial buffering provided in an embodiment of the present application, which shows a process of spatially superimposing a plurality of circular buffer zones;

[0196] Figure 4 It is a schematic diagram of the result of superposition value accumulation provided in an embodiment of the present application, which shows the result of superposition value accumulation obtained after spatial superposition of several circular buffer zones.

[0197] The candidate unit of this embodiment divides the abnormal area set according to the numerical value of the coefficient of variation, and can objectively and accurately identify areas with different abnormal degrees and classify them; randomly extracting several abnormal areas from the divided abnormal area categories as data samples not only ensures the representativeness of the samples, but also avoids the subjectivity of human selection. The subclass grid in the data sample is simplified into a uniform dot matrix, which simplifies the complexity of the data and improves the processing efficiency. Taking each point in the dot matrix set as the center of the circle and establishing a circular buffer according to a preset radius can intuitively reflect the spatial characteristics around each point for spatial superposition and analysis.

[0198] The take-off and landing unit is used to obtain detailed road network data within a preset area, including information such as road geometry, type and traffic restrictions;

[0199] The take-off and landing unit is also used to overlay the location of the best airport placement candidate area with the road network data to determine the optimal drone take-off point;

[0200] The take-off and landing unit is also used to use an efficient path search algorithm (such as Dijkstra algorithm) combined with the UAV flight path planned using the minimum circle cover algorithm to calculate the shortest path from the selected optimal UAV take-off point to all destination sample points;

[0201] The take-off and landing unit is also used to determine the final layout point of the drone airport based on the calculation results of the shortest path and the actual operation requirements of the drone, taking into account factors such as the location of the take-off point, the surrounding environment and traffic conditions, so as to obtain the take-off and landing point of the drone;

[0202] The take-off and landing unit is also used to set up fixed drone airports and mobile drone airports based on the take-off and landing points of drones; and ensure that the facilities of the drone airports are complete and can meet the needs of drone take-off and landing and maintenance; among them, fixed drone airports are set up in preset key supervision areas, and mobile drone airports are set according to actual needs.

[0203] The acquisition unit is used to calibrate the satellite's multispectral camera with radiation correction plates of different reflectivity of 25%, 50% and 75% before the satellite passes overhead to ensure its stable performance and high accuracy; and record the calibration results for subsequent data processing and analysis; wherein the satellite is a multispectral satellite;

[0204] The acquisition unit is also used to control the UAV to depart from the UAV airport (take-off and landing point) and arrive at the destination sample point; adjust the observation angle (azimuth) and camera parameters (camera pitch angle) of the UAV to be consistent with the observation conditions of the data acquired by the satellite;

[0205] The acquisition unit is also used to control the UAV and satellite to simultaneously acquire image data of the project supervision area, thereby obtaining UAV data and high-resolution satellite remote sensing data; wherein the image data is multispectral image data;

[0206] The acquisition unit is also used to perform supplementary photography or encrypted measurements through drones if there are areas where satellite remote sensing data is missing or where the emissivity changes exceed a preset value, to ensure the integrity and accuracy of the data.

[0207] In the acquisition unit of this embodiment, satellite remote sensing data has the advantages of macroscopicity, continuity and timeliness, but may be missing or have reduced quality due to factors such as cloud cover and ground cover; and drones can flexibly adjust flight routes and altitudes to perform supplementary photography or encrypted measurements of specific areas, thereby filling gaps in satellite data or improving data accuracy.

[0208] In one embodiment, the correction module 20 is specifically:

[0209] Perform preliminary preprocessing on drone data, such as denoising and enhancement, and extract reflectivity information from drone data for subsequent data analysis;

[0210] Time-synchronize drone data with satellite remote sensing data to ensure data consistency and comparability;

[0211] Based on the actual radiation reflectivity of each band in the UAV data, the satellite remote sensing data is corrected to eliminate the differences caused by atmospheric correction and obtain multispectral reflectivity data; the accuracy and reliability of the multispectral reflectivity data are evaluated;

[0212] The relevant index (such as NDVI) is calculated based on the multispectral reflectance data, and the latest relevant index is used as the basis for the next sampling.

[0213] In one embodiment, the evaluation module 30 includes a supervision unit and a feedback unit; wherein the supervision unit is a process of obtaining the supervision results of the ecological restoration project, and the feedback unit is a process of continuous monitoring and feedback, specifically:

[0214] The supervision unit is used to conduct time series analysis and spatial distribution analysis on the NDVI data after time series fusion in the multispectral reflectance data, so as to evaluate the ecological restoration effect of the project supervision area and obtain the first supervision result;

[0215] The supervision unit is also used to combine the first supervision result with the field investigation results of the project supervision area and preset expert knowledge to identify existing problems and challenges and obtain the supervision result of the ecological restoration project.

[0216] Feedback unit, used to formulate regular monitoring plans to ensure continuous monitoring of the project supervision area; conduct irregular spot checks on the project supervision area according to actual needs and monitoring results; dynamically adjust sampling plans and supervision priorities according to monitoring results and the progress of ecological restoration projects;

[0217] The feedback unit is also used to establish a real-time data transmission system to transmit monitoring data to the supervision system in real time; use data visualization technology to convert monitoring data into intuitive charts and images; and use visualization to display the progress and results of ecological restoration projects in real time;

[0218] The feedback unit is also used to establish an early warning system, set early warning thresholds and trigger conditions; monitor and analyze monitoring data in real time to identify abnormal situations; once an abnormal situation is found, a warning signal is issued in a timely manner and a response is carried out;

[0219] The feedback unit is also used to collect and analyze monitoring data, evaluate the progress and effectiveness of ecological restoration projects; regularly prepare monitoring reports to summarize monitoring results and problems found; and submit monitoring reports to project management departments and relevant policy-making agencies to provide a scientific basis for project management and policy making.

[0220] Overall, this embodiment has the following beneficial effects:

[0221] In the present invention, the high-resolution multispectral image data of the UAV can clearly reflect the surface details and help to find potential problem areas; and the satellite remote sensing data can reveal the changing trends of surface features through time series analysis; by combining the data of the two, they can verify and correct each other to improve the accuracy and reliability of the data. The UAV has the ability to be quickly deployed and monitored in real time, and can monitor key areas multiple times in a short period of time, thereby improving monitoring efficiency; and although the acquisition cycle of satellite remote sensing data is relatively long, once acquired, it can cover a large area and reduce duplication of work; the combination of the two makes monitoring more efficient. In addition, by identifying ecological change trends and potential problem areas, it is possible to accurately determine the areas where the UAV needs to focus, so that take-off and landing points can be set in these areas. This precise positioning reduces aimless and large-scale flights, so the data collection capabilities of the UAV can be maximized to ensure that key areas are quickly monitored and evaluated;

[0222] Through the deep integration and correction technology of drone and satellite remote sensing data, full-cycle, high-precision, and low-cost supervision of landscape and ecological restoration projects has been achieved. This method not only significantly improves the accuracy and timeliness of data, but also effectively reduces the cost of human verification and enhances the intelligence and automation level of supervision. At the same time, through continuous monitoring and feedback mechanisms, problems arising in the ecological restoration process can be discovered and resolved in a timely manner to ensure the smooth implementation of the project and achieve the expected results.

[0223] Embodiment three:

[0224] The embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the method for supervising ecological restoration projects based on drones;

[0225] Wherein, the ecological restoration engineering supervision method based on drones, if implemented in the form of a software functional unit and used as an independent product, can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned method embodiments when executed by the processor. Wherein, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium, etc.

[0226] The above are preferred embodiments of the present invention. It should be noted that, for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for monitoring ecological restoration projects based on drones, characterized in that: include: Controlling the drone and the satellite to simultaneously collect image data of the project supervision area, and obtaining drone data and satellite remote sensing data; wherein the take-off and landing points of the drone are determined by identifying the ecological change trend and potential problem areas of the project supervision area; Correcting the satellite remote sensing data based on the drone data to obtain multispectral reflectance data; The ecological restoration effect of the project supervision area is evaluated based on the multispectral reflectance data to obtain the ecological restoration project supervision results.

2. The method for monitoring ecological restoration projects based on drones according to claim 1, characterized in that: The take-off and landing points of the drone are determined by identifying the ecological change trends and potential problem areas in the project supervision area, specifically: Dividing the project supervision area into a two-dimensional grid; By performing curve compression and clustering processing on the two-dimensional grid, the coefficient of variation of the subclass grids in the two-dimensional grid is calculated; The subclass grids in the two-dimensional grid whose coefficient of variation exceeds a preset threshold are defined as abnormal regions, to obtain an abnormal region set; The take-off and landing points of the UAV are obtained by performing spatial superposition processing on the abnormal areas in the abnormal area concentration.

3. The method for monitoring ecological restoration projects based on drones as claimed in claim 2, characterized in that: By performing curve compression and clustering processing on the two-dimensional grid, the coefficient of variation of the subclass grids in the two-dimensional grid is calculated, specifically: Obtaining the normalized vegetation index of the subclass grid from the two-dimensional grid to obtain initial time series data; Performing curve compression on the initial time series data to obtain first time series data; Based on the first set of time series data points in the first time series data, connecting the first point and the last point into a straight line to obtain a target straight line; Calculating the distances from all points in the first set of time series data points to the target straight line to obtain a distance set; Based on the maximum value in the distance set, target data of the time series data is eliminated to obtain thinned second time series data; The coefficient of variation of the subclass grid in the two-dimensional grid is calculated based on the second time series data.

4. The method for monitoring ecological restoration projects based on drones as claimed in claim 3, characterized in that: Based on the maximum value in the distance set, the target data of the time series data is eliminated to obtain the thinned second time series data, specifically: Obtaining the maximum value in the distance set to obtain a maximum distance value; If the maximum distance value is less than a preset threshold, other points except the first and last points in the time series curve are removed; wherein the time series curve is composed of the time series data; If the maximum distance value is greater than or equal to the preset threshold, retaining the first data point corresponding to the maximum distance value, and dividing the first group of time series data points into two sub-point sets with the first data point as the boundary; The two sub-point sets are processed recursively, and the thinned second time series data is obtained by traversing a plurality of time series data point sets in the time series data.

5. The method for monitoring ecological restoration projects based on drones as claimed in claim 3, characterized in that: The coefficient of variation of the subclass grid in the two-dimensional grid is calculated according to the second time series data, specifically: Controlling the subclass grids in the second time series data to cluster with the subclass grids within a preset radius to obtain a clustering result; Based on the values ​​of preset time points in the clustering results, spatially clustering the data of several categories in the clustering results to obtain a comprehensive subclass grid set; The coefficient of variation of the comprehensive subclass grid set in the two-dimensional grid is obtained by calculating the degree of discreteness of the data distribution of the comprehensive subclass grid set.

6. The method for monitoring ecological restoration projects based on drones according to claim 2, characterized in that: By performing spatial superposition processing on the abnormal areas in the abnormal area concentration, the take-off and landing points of the UAV are obtained, specifically: According to the numerical value of the coefficient of variation, the abnormal region set is divided into a number of abnormal region categories; Randomly select a number of abnormal regions from the several abnormal region categories to obtain data samples; Simplifying the subclass grids in the data sample into a uniform dot matrix to obtain a dot matrix set; Taking each point in the dot matrix as the center of a circle, a plurality of circular buffer zones are established according to a preset radius; The plurality of circular buffer zones are spatially superimposed to obtain a numerical distribution of preset objective function values ​​corresponding to each zone; Based on the numerical distribution, several areas with the highest values ​​are selected as the take-off and landing points of the UAV.

7. The method for monitoring ecological restoration projects based on drones according to claim 1, characterized in that: Control the drone and satellite to collect image data of the project supervision area at the same time, and obtain drone data and satellite remote sensing data, specifically: Adjusting the observation angle and camera parameters of the drone to be consistent with the observation conditions of the data acquired by the satellite; Controlling the UAV and the satellite to simultaneously collect image data of the project supervision area to obtain the UAV data and the satellite remote sensing data; If there are areas where data is missing or where the emissivity changes exceed a preset value in the satellite remote sensing data, supplementary photography or encrypted measurement is performed by the drone.

8. The method for monitoring ecological restoration projects based on drones as claimed in claim 1, characterized in that: The project supervision area is specifically: Obtain basic geographic information of the comprehensive project supervision area; Based on the basic geographic information, by identifying the distribution of different land types and soil types, the key monitoring areas in the comprehensive project supervision area are determined to obtain the project supervision area.

9. An ecological restoration project monitoring device based on drones, characterized in that: It includes data module, correction module and evaluation module; The data module is used to control the drone and the satellite to simultaneously collect image data of the project supervision area, thereby obtaining drone data and satellite remote sensing data; wherein the take-off and landing points of the drone are determined by identifying the ecological change trends and potential problem areas of the project supervision area; The correction module is used to correct the satellite remote sensing data based on the UAV data to obtain multispectral reflectance data; The evaluation module is used to evaluate the ecological restoration effect of the project supervision area according to the multispectral reflectance data to obtain the ecological restoration project supervision results.

10. A storage medium, characterized in that: The storage medium stores a computer program, which is called and executed by a computer to implement a drone-based ecological restoration project supervision method as described in any one of claims 1 to 8.

Images