Supplementary cultivated land recovery evaluation method and system and medium

Abstract

The invention belongs to the field of supplementary cultivated land detection and evaluation, and particularly relates to a supplementary cultivated land recovery evaluation method and system and a medium, which utilize the advantages of flexibility and short acquisition period of an unmanned aerial vehicle to realize accurate mapping of geographic coordinates of multi-source data collaboration, and specifically perform SIFT feature matching algorithm on an unmanned aerial vehicle image and a high-resolution remote sensing image so as to improve the accuracy of the supplementary cultivated land recovery evaluation. The problem of pixel coordinate geography conversion of the unmanned aerial vehicle image is solved, the supplementary cultivated land pattern spots are mapped to the unmanned aerial vehicle image, and the accuracy of cultivated land range checking is ensured. Meanwhile, instance segmentation is carried out on a cultivated land pattern spot range mapped by the unmanned aerial vehicle image by using a YOLO algorithm, cultivated land and non-cultivated land areas in geographical pattern spots are accurately distinguished, and a solid data support is provided for quantitative evaluation of the cultivated land area. The method effectively solves the long-standing problems of time lag, insufficient precision, high cost and other pain points in supplemented cultivated land supervision, and provides reliable technical support for the intelligent monitoring of the supplemented cultivated land.

Description

Technical Field

[0001] This application relates to the technical field of supplementary cultivated land detection and evaluation, and particularly to a method, system and medium for evaluating the restoration of supplementary cultivated land. Background Art

[0002] Traditional supervision methods for supplementary cultivated land mainly rely on manual inspection and spot checks, which have disadvantages such as long cycle, low efficiency, and being easily interfered by human factors, and cannot meet the needs of modern supervision. In recent years, remote sensing image interpretation has been widely used in land type segmentation and recognition, but there are still many limitations in actual applications. On the one hand, the acquisition cost of commercial remote sensing images is relatively high, and due to the long acquisition cycle, the update frequency cannot meet the needs of real-time monitoring; on the other hand, remote sensing images are also easily affected by factors such as weather, season, and light. In addition, remote sensing image interpretation usually involves the processing of large-area data, with a large amount of calculation and a long interpretation time, making it difficult to meet the needs of rapid supervision. These factors may lead to a decrease in the stability and consistency of images, thus affecting the accuracy of cultivated land segmentation.

[0003] In contrast, drones have more flexibility in cultivated land monitoring due to their advantages of low cost, high frequency, and high resolution, and are suitable for dynamic monitoring of cultivated land. However, at present, the inspection of supplementary cultivated land by drones only involves the backhaul of video streams and still relies on manual visual analysis, without fully combining intelligent algorithms for automatic recognition and analysis, making it difficult to efficiently detect problems of false supplementary cultivated land. Therefore, how to integrate deep learning and computer vision to transform the drone inspection from passive monitoring to intelligent analysis is the key to improving supervision efficiency. Summary of the Invention

[0004] Embodiments of this application provide a method, system and medium for evaluating the restoration of supplementary cultivated land, which uses drones as the carrier and combines feature matching and segmentation algorithms to achieve precise identification and dynamic monitoring of supplementary cultivated land, improving the timeliness and flexibility of monitoring to solve problems such as monitoring lag and low efficiency existing in current supplementary cultivated land supervision.

[0005] To this end, according to one aspect of this application, a method for evaluating the restoration of supplementary cultivated land is provided, including the following steps: Obtain the vector boundary layer of supplementary cultivated land and high-resolution remote sensing images; Configure the drone and navigate to the geometric center coordinates of each supplementary cultivated land patch to collect drone images of the supplementary cultivated land area; Use the SIFT feature matching algorithm to extract the feature points of the high-resolution remote sensing image and the drone image, and calculate and evaluate the homography matrix; Perform perspective transformation using the homography matrix to map the supplementary cultivated land patch of the high-resolution remote sensing image into the drone image; Segment the cultivated land and non-cultivated land areas within the supplementary cultivated land patches based on the YOLO algorithm, and calculate their geographical areas and proportions respectively through the homography matrix.

[0006] Optionally, obtaining the vector boundary layer of the supplementary cultivated land and the high-resolution remote sensing image specifically includes the following steps: Obtain the vector boundary layer of the supplementary cultivated land from the cultivated land business system, and obtain the high-resolution remote sensing image with a resolution better than 1 meter, and the size of the high-resolution remote sensing image covers the entire vector boundary layer of the supplementary cultivated land; if the coordinates of the vector boundary layer of the supplementary cultivated land are inconsistent with those of the high-resolution remote sensing image, perform coordinate transformation on the vector boundary layer to make its coordinate system consistent with that of the high-resolution remote sensing image; Overlay the vector boundary layer of the supplementary cultivated land on the high-resolution remote sensing image to determine the spatial position and range of each supplementary cultivated land patch; Find no less than 5 landmark points of geographical coordinates around each supplementary cultivated land patch, and crop the image according to the position of each supplementary cultivated land patch to ensure that the cropped image only contains the area of the cultivated land patch and the surrounding landmark points.

[0007] Optionally, collecting the UAV images of the supplementary cultivated land area specifically includes: Determine the geometric center coordinates of the supplementary cultivated land patch; Control the UAV to navigate to the geometric center of the cultivated land patch and collect the UAV image; The UAV image completely covers the boundary of the supplementary cultivated land, and ensure that the pixel landmark points corresponding to no less than 5 geographical coordinates selected in the high-resolution remote sensing image can be identified in the UAV image.

[0008] Optionally, using the SIFT feature matching algorithm to extract the feature points of the high-resolution remote sensing image and the UAV image, and calculating and evaluating the homography matrix specifically includes the following steps: Convert the UAV image and the high-resolution remote sensing image into grayscale images and perform Gaussian blur denoising; Detect the feature points in the UAV image and the high-resolution remote sensing image based on the SIFT algorithm, and the feature points include corner points and edges; Use Lowe’s ratio test to screen the matching points. The matching points are the pairs of points with similar features found in the UAV image and the high-resolution remote sensing image through the SIFT algorithm, representing the corresponding positions in the images; By calculating the ratio of the distance between the closest matching point and the second closest matching point of each feature point, the calculation formula is: , where is the distance between the feature point and the closest matching point, It is the distance between the feature point and the second closest matching point. If this ratio is lower than the set threshold, then this matching point pair is retained; Use the FLANN algorithm to match the feature points in the UAV image and the high-resolution remote sensing image by calculating the Euclidean distance between the descriptors of the feature points; Based on the RANSAC algorithm, iteratively optimize the matching point pairs and eliminate the mismatched points; Convert the pixel coordinates of the matching points in the high-resolution remote sensing image into geographic coordinates; Utilize the pixel coordinates of the UAV image matching points and the geographic coordinates of the high-resolution remote sensing image matching points, and fit the homography transformation matrix between the image and the image through the least squares method; Calculate the reprojection error of the homography matrix using the identification points with known geographic coordinates in the UAV image. If the root mean square error of the reprojection errors of all identification points is greater than 10 pixels, optimize the matching points and update the homography matrix.

[0009] Optionally, optimize the matching points and update the homography matrix. Specifically, by adjusting the threshold of the Lowe’s ratio test, optimizing the nearest neighbor search parameters of the FLANN algorithm, and improving the mismatched point elimination strategy of the RANSAC algorithm, optimize the SIFT matching points to obtain the optimal homography matrix.

[0010] Optionally, use the homography matrix for perspective transformation to map the supplementary cultivated land patches in the high-resolution remote sensing image to the UAV image, which specifically includes: Based on the homography matrix, perform perspective transformation on each boundary point of the supplementary cultivated land patches in the high-resolution remote sensing image, map the vector boundary layer coordinates to the UAV image coordinate system, and achieve the precise registration of geographic space coordinates and image pixel coordinates.

[0011] Optionally, based on the YOLO algorithm, segment the cultivated land and non-cultivated land areas within the supplementary cultivated land patches, and calculate their geographic areas and proportions respectively through the homography matrix, which specifically includes the following steps: After overlaying the supplementary cultivated land patches on the UAV image, perform semantic segmentation of the cultivated land and non-cultivated land on the UAV image within the patch range, and output the pixel-level segmentation result; Convert the polygon image coordinates of the segmented cultivated land and non-cultivated land into geographic coordinates through the inverse matrix of the homography matrix, and calculate the actual areas of the cultivated land and non-cultivated land in the supplementary plot; Calculate the area ratio of the cultivated land and non-cultivated land in the geographic coordinate system, and generate an assessment report on the restoration progress of the supplementary cultivated land.

[0012] According to another aspect of the present application, a supplementary cultivated land restoration assessment system is provided, including: The multi-source data acquisition module is used to access high-resolution remote sensing image data, overlay and supplement the cultivated land area range; and plan the UAV flight path, trigger the image acquisition instruction to ensure overlap with the high-resolution remote sensing image. The image registration module performs denoising and gray normalization operations to unify the imaging conditions of the high-resolution remote sensing image and the UAV image. It generates descriptors through SIFT multi-scale feature detection, combines FLANN matching and RANSAC optimization to calculate the homography matrix, and loads the homography matrix to project the supplementary cultivated land patches in the high-resolution remote sensing image onto the UAV image. The cultivated land segmentation module uses the YOLO algorithm to segment the cultivated land and non-cultivated land pixel regions within the cultivated land patches, calculates the cultivated land coverage rate and the proportion of non-cultivated land, and outputs quantitative indicators. The evaluation report generation module is used to generate an evaluation report on supplementary cultivated land with maps.

[0013] According to another aspect of the present application, there is provided a computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned supplementary cultivated land restoration evaluation method are implemented.

[0014] The beneficial effects of the supplementary cultivated land restoration evaluation method, system and medium provided by the present application are as follows: Taking advantage of the flexibility and short acquisition cycle of UAVs, it realizes the precise mapping of geographical coordinates for multi-source data collaboration. Specifically, it performs SIFT feature matching algorithm on UAV images and high-resolution remote sensing images, solves the geographical conversion of pixel coordinates of UAV images, maps the supplementary cultivated land patches onto the UAV images, and ensures the accuracy of cultivated land area verification. At the same time, the YOLO algorithm is used to perform instance segmentation on the mapped cultivated land patch range of UAV images, accurately distinguish the cultivated land and non-cultivated land areas within the geographical patches, and provide solid data support for the quantitative evaluation of cultivated land area. This solution constructs a full-chain solution covering data acquisition, processing, and decision support through a technical loop of "precise coordinate mapping - intelligent image analysis - dynamic monitoring and early warning", effectively solving the long-existing pain points such as time lag, insufficient accuracy, and high cost in the supervision of supplementary cultivated land, and providing reliable technical support for the intelligent monitoring of supplementary cultivated land. Description of the Drawings

[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required to be used in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0016] Among them: Figure 1 It is a flowchart of a supplementary cultivated land restoration evaluation method shown in an embodiment of the present application; Figure 2 It is a schematic structural diagram of a supplementary cultivated land restoration evaluation system shown in an embodiment of the present application. Detailed implementation manners

[0017] To facilitate the understanding of the present application, the present application will be described more comprehensively below with reference to the relevant drawings. The preferred embodiments of the present application are given in the drawings. However, the present application can be implemented in many other different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure content of the present application more thorough and comprehensive.

[0018] In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the present application, "a plurality" means two or more unless otherwise specifically defined.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which this application belongs. The terms used in the description of the present application herein are only for the purpose of describing specific embodiments and are not intended to limit the present application.

[0020] According to one aspect of the present application, an embodiment of the present application provides a supplementary cultivated land restoration evaluation method, as Figure 1 shown, the supplementary cultivated land restoration evaluation method includes the following steps: S1. Obtain the vector boundary layer of the supplementary cultivated land area and the high-resolution remote sensing image.

[0021] Specifically, step S1 includes the following steps: Obtain the vector boundary layer of the supplementary cultivated land from the cultivated land business system, and obtain the high-resolution remote sensing image with a resolution better than 1 meter, and the size of the high-resolution remote sensing image should cover the entire vector boundary layer of the supplementary cultivated land; if the coordinates of the vector boundary layer of the supplementary cultivated land are inconsistent with those of the high-resolution remote sensing image, the seven-parameter transformation method is used to perform coordinate transformation on the vector boundary layer to make it unified with the coordinate system (CGCS2000) of the remote sensing image; Overlay the vector boundary layer of the supplementary cultivated land on the remote sensing image to determine the spatial position and scope of each supplementary cultivated land patch; Find no less than 5 landmark points with geographical coordinates around each supplementary cultivated land patch, and crop the image according to the position of each supplementary cultivated land patch to ensure that the cropped image only contains the area of the patch and the surrounding landmark points.

[0022] S2. Configure the drone and navigate it to the geometric center coordinates of the supplementary cultivated land patch to collect drone images of the supplementary cultivated land area.

[0023] Specifically, step S2 includes the following steps: Determine the geometric center coordinates of the supplementary cultivated land patch; Control the drone to navigate to the geometric center of the cultivated land patch and collect drone images; The drone images should completely cover the boundary of the supplementary cultivated land and ensure that the pixel landmark points corresponding to no less than 5 geographical coordinates selected in the remote sensing image can be identified in the drone images.

[0024] S3. Use the SIFT feature matching algorithm to extract the feature points of the high-resolution remote sensing image and the drone image, and calculate and evaluate the homography matrix.

[0025] Specifically, step S3 includes the following steps: Convert the drone image and the high-resolution remote sensing image into grayscale images, and perform Gaussian blur denoising using a 5×5 Gaussian kernel; Detect significant feature points such as corners and edges in the drone image and the high-resolution remote sensing image based on the SIFT algorithm; Use Lowe’s ratio test to screen the matching points. The matching points refer to the pairs of points with similar features found in the drone image and the high-resolution remote sensing image through the SIFT algorithm, representing the corresponding positions in the images; By calculating the ratio of the distance between the closest matching point and the second-closest matching point for each feature point, the calculation formula is: , where is the distance between the feature point and the closest matching point, is the distance between the feature point and the second-closest matching point. If the ratio is lower than the set threshold, then retain the pair of matching points (i.e., the above-mentioned closest matching point and the second-closest matching point); it should be noted here that the set threshold should be lower than 0.7, aiming to balance the accuracy and the number of matching points and effectively eliminate false matches; Use the FLANN algorithm to match the feature points in the two images by calculating the Euclidean distance between the feature descriptors; Based on the RANSAC algorithm, iteratively optimize the pairs of matching points to eliminate false matching points; Convert the pixel coordinates of the matching points of the high-resolution remote sensing image into geographical coordinates; Using the pixel coordinates of the matching points in the UAV image and the geographic coordinates of the matching points in the high-resolution remote sensing image, the homography transformation matrix between the image and the image is fitted by the least squares method. Specifically, the homography matrix H can be calculated through the following relationship:

[0026] where (x, y) are the geographic coordinates of the remote sensing image, (x', y') are the pixel coordinates of the UAV image, w' is the weight factor, and H is the required homography matrix; Calculate the reprojection error of the homography matrix using the identification points with known geographic coordinates in the UAV image. If the root mean square error of the reprojection errors of all identification points is greater than 10 pixels, optimize the matching points and update the homography matrix. Optimize the matching points and update the homography matrix by adjusting the threshold of the Lowe’s ratio test, optimizing the nearest neighbor search parameters of the FLANN algorithm, and improving the mis-matching rejection strategy of the RANSAC algorithm to optimize the SIFT matching points to obtain the optimal homography matrix.

[0027] S4. Use the homography matrix for perspective transformation to map the supplementary cultivated land patches in the high-resolution remote sensing image into the UAV image; Specifically, step S4 includes: performing perspective transformation on each boundary point of the supplementary cultivated land patches in the high-resolution remote sensing image based on the homography matrix H, mapping the vector boundary layer coordinates (x, y) to the UAV image coordinate system (x', y'), and realizing the accurate registration of geographic space coordinates and image pixel coordinates.

[0028] S5. Based on the YOLOv11 algorithm, segment the cultivated land and non-cultivated land areas within the supplementary cultivated land patches, and calculate their geographic areas and proportions respectively through the homography matrix; Specifically, step S5 includes the following steps: After overlaying the supplementary cultivated land patches on the UAV image, perform semantic segmentation of the cultivated land and non-cultivated land on the UAV image within the patch range, and output the pixel-level segmentation result; Convert the polygon image coordinates of the segmented cultivated land and non-cultivated land into geographic coordinates through the inverse matrix of the homography matrix H, and calculate the actual areas of the cultivated land and non-cultivated land in the supplementary plot; among them, the calculation formula for converting the point (x', y') in the image coordinate system to the point (x, y) in the geographic coordinate system is as follows:

[0029] Calculate the area ratio of the cultivated land and non-cultivated land in the geographic coordinate system, and generate an evaluation report on the restoration progress of the supplementary cultivated land.

[0030] In the supplementary cultivated land restoration assessment method in the embodiments of the present application, by leveraging the advantages of drones, such as flexibility and short acquisition cycle, precise mapping of geographical coordinates with multi-source data collaboration is achieved. Specifically, the SIFT feature matching algorithm is used for the drone images and high-resolution remote sensing images to solve the geographical transformation of the pixel coordinates of the drone images, map the supplementary cultivated land patches onto the drone images, and ensure the accuracy of cultivated land scope verification. At the same time, the YOLO algorithm is used to perform instance segmentation on the cultivated land patch ranges mapped by the drone images, accurately distinguish the cultivated land and non-cultivated land areas within the geographical patches, and provide solid data support for the quantitative assessment of the cultivated land area. Through the technical closed-loop of "precise coordinate mapping - intelligent image analysis - dynamic monitoring and early warning", this solution constructs a full-chain solution covering data acquisition, processing, and decision support, effectively solving the long-existing pain points in supplementary cultivated land supervision, such as time lag, insufficient accuracy, and high cost, and providing reliable technical support for the intelligent monitoring of supplementary cultivated land.

[0031] According to another aspect of the present application, embodiments of the present application also provide a supplementary cultivated land restoration assessment system, as Figure 2 shown, including: A multi-source data acquisition module, used to access high-resolution remote sensing image data, overlay the supplementary cultivated land area range; and plan the drone flight path, trigger the image acquisition instruction, and ensure overlap with the high-resolution remote sensing image; An image registration module, which performs denoising and gray-scale normalization operations to unify the imaging conditions of the high-resolution remote sensing image and the drone image, generates descriptors through SIFT multi-scale feature detection, combines FLANN matching and RANSAC optimization to calculate the homography matrix, and loads the homography matrix to project the supplementary cultivated land patches in the high-resolution remote sensing image onto the drone image; A cultivated land segmentation module, which uses the YOLO algorithm to segment the cultivated land and non-cultivated land pixel areas within the cultivated land patches, statistically calculates the cultivated land coverage rate and the proportion of non-cultivated land, and outputs quantitative indicators; An evaluation report generation module, used to generate a supplementary cultivated land evaluation report with drawings.

[0032] In addition, embodiments of the present application also disclose a computer-readable storage medium.

[0033] Specifically, computer-readable instructions are stored in the computer-readable medium. When the computer-readable instructions are executed by one or more processors, the one or more processors are caused to execute the steps of the supplementary cultivated land restoration assessment method.

[0034] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this specification.

[0035] The above embodiments only represent several implementation manners of the present application, and the description thereof is relatively specific and detailed. However, it should not be construed as a limitation on the scope of the application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the appended claims.

Claims

1. A method for evaluating the restoration of cultivated land for supplementary purposes, characterized in that, Including the following steps: Obtain the vector boundary layer of the supplementary cultivated land and the high-resolution remote sensing image; Configure the drone and navigate it to the geometric center coordinates of each supplementary cultivated land patch, and collect the drone images of the supplementary cultivated land area; Use the SIFT feature matching algorithm to extract the feature points of the high-resolution remote sensing image and the drone image, and calculate and evaluate the homography matrix; Perform perspective transformation using the homography matrix to map the supplementary cultivated land patches of the high-resolution remote sensing image into the drone image; Based on the YOLO algorithm, segment the cultivated land and non-cultivated land areas within the supplementary cultivated land patches, and calculate their geographical areas and proportions respectively through the homography matrix.

2. The supplementary cultivated land restoration assessment method according to claim 1, wherein The specific steps for obtaining the vector boundary layer of the supplementary cultivated land and the high-resolution remote sensing image include the following: Obtain the vector boundary layer of the supplementary cultivated land from the cultivated land business system, and obtain the high-resolution remote sensing image with a resolution better than 1 meter. The size of the high-resolution remote sensing image covers the entire vector boundary layer of the supplementary cultivated land; if the coordinate systems of the vector boundary layer of the supplementary cultivated land and the high-resolution remote sensing image are inconsistent, perform coordinate transformation on the vector boundary layer to make it unified with the coordinate system of the high-resolution remote sensing image; Overlay the vector boundary layer of the supplementary cultivated land on the high-resolution remote sensing image to determine the spatial position and range of each supplementary cultivated land patch; Find no less than 5 identification points of geographical coordinates around each supplementary cultivated land patch, and crop the image according to the position of each supplementary cultivated land patch to ensure that the cropped image only contains the area of the cultivated land patch and the identification points around it.

3. The supplementary cultivated land restoration assessment method according to claim 2, wherein The specific steps for collecting the drone images of the supplementary cultivated land area include: Determine the geometric center coordinates of the supplementary cultivated land patch; Control the drone to navigate to the geometric center of the cultivated land patch and collect the drone image; The drone image completely covers the boundary of the supplementary cultivated land, and ensures that the pixel identification points corresponding to no less than 5 geographical coordinates selected in the high-resolution remote sensing image can be identified in the drone image.

4. The supplementary cultivated land restoration assessment method according to claim 1, wherein The specific steps for using the SIFT feature matching algorithm to extract the feature points of the high-resolution remote sensing image and the drone image, and calculate and evaluate the homography matrix include the following: Convert the drone image and the high-resolution remote sensing image into grayscale images and perform Gaussian blur denoising; Detect the feature points in the drone image and the high-resolution remote sensing image based on the SIFT algorithm. The feature points include corner points and edges; Use the Lowe’s ratio test to screen the matching points. The matching points are pairs of points with similar features found in the drone image and the high-resolution remote sensing image through the SIFT algorithm, representing the corresponding positions in the images; By calculating the ratio of the distance between the closest matching point and the second-closest matching point for each feature point, the calculation formula is: , where is the distance between the feature point and the closest matching point, is the distance between the feature point and the second-closest matching point. If the ratio is lower than the set threshold, then retain the pair of matching points; Use the FLANN algorithm to match the feature points in the drone image and the high-resolution remote sensing image by calculating the Euclidean distance between the descriptors of the feature points; Based on the RANSAC algorithm, perform iterative optimization on the matching point pairs to eliminate the mis-matching points; Convert the pixel coordinates of the matching points of the high-resolution remote sensing image into geographical coordinates; Utilize the pixel coordinates of the matching points of the drone image and the geographical coordinates of the matching points of the high-resolution remote sensing image, and fit the homography transformation matrix between the image and the remote sensing image by the least squares method. Calculate the reprojection error of the homography matrix using the landmark points with known geographic coordinates in the UAV images. If the root mean square error of the reprojection errors of all landmark points is greater than 10 pixels, optimize the matching points and update the homography matrix.

5. The supplementary cultivated land restoration assessment method according to claim 4, wherein The optimization of the matching points and the update of the homography matrix are specifically achieved by adjusting the threshold of the Lowe’s ratio test, optimizing the nearest neighbor search parameters of the FLANN algorithm, and improving the mis - matching rejection strategy of the RANSAC algorithm to optimize the SIFT matching points and obtain the optimal homography matrix.

6. The supplementary cultivated land restoration assessment method according to claim 1, characterized in that The use of the homography matrix for perspective transformation to map the supplementary cultivated land patches in the high - resolution remote sensing image into the UAV image specifically includes: Based on the homography matrix, perform perspective transformation on each boundary point of the supplementary cultivated land patches in the high - resolution remote sensing image, map the vector boundary layer coordinates to the UAV image coordinate system, and achieve the accurate registration of geographic space coordinates and image pixel coordinates.

7. The supplementary cultivated land restoration assessment method according to claim 1, wherein The segmentation of cultivated land and non - cultivated land areas within the supplementary cultivated land patches based on the YOLO algorithm and the calculation of their geographic areas and proportions through the homography matrix specifically include the following steps: After overlaying the supplementary cultivated land patches on the UAV image, perform semantic segmentation of cultivated land and non - cultivated land on the UAV image within the patch range and output the pixel - level segmentation result; Convert the polygon image coordinates of the segmented cultivated land and non - cultivated land into geographic coordinates through the inverse matrix of the homography matrix, and calculate the actual areas of cultivated land and non - cultivated land in the supplementary plot; Calculate the area ratio of cultivated land and non - cultivated land in the geographic coordinate system and generate an evaluation report on the restoration progress of the supplementary cultivated land.

8. A supplementary cultivated land restoration assessment system, characterized in that, It includes: A multi - source data acquisition module for accessing high - resolution remote sensing image data and overlaying the supplementary cultivated land area range; And planning the UAV flight path, triggering the image acquisition instruction to ensure overlap with the high - resolution remote sensing image; An image registration module that performs denoising and gray - scale normalization operations to unify the imaging conditions of the high - resolution remote sensing image and the UAV image, generates descriptors through SIFT multi - scale feature detection, combines FLANN matching and RANSAC optimization to calculate the homography matrix, and loads the homography matrix to project the supplementary cultivated land patches in the high - resolution remote sensing image onto the UAV image; A cultivated land segmentation module that uses the YOLO algorithm to segment the cultivated land and non - cultivated land pixel areas within the cultivated land patch, statistically calculates the cultivated land coverage rate and the proportion of non - cultivated land, and outputs quantitative indicators; An evaluation report generation module for generating an evaluation report on the supplementary cultivated land with graphics.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the supplementary cultivated land restoration evaluation method according to any one of claims 1 - 7.

Images