Performance evaluation method and system for natural resource financial project

Abstract

The invention discloses a natural resource financial project performance evaluation method and system, and relates to the technical field of engineering supervision. Decomposing the ecological restoration indexes to obtain process targets in different stages; according to high-altitude remote sensing and low-altitude remote sensing, the project construction process is monitored, and the ecological restoration area of the current stage is obtained; the ecological restoration area comprises a newly increased forest land area and a newly increased cultivated land area; and comparing the ecological restoration area of the current stage with a process target, and generating a performance report of the current stage. Monitoring is conducted through the high-altitude remote sensing technology and the low-altitude remote sensing technology, ecological restoration data in a wide area can be rapidly obtained, the low-efficiency process of manual access is avoided, and the data collection efficiency is greatly improved. Quantitative analysis is carried out through remote sensing data, errors possibly caused by manual data filling and reporting are avoided, and therefore the accuracy of performance evaluation is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of engineering supervision, and in particular to a natural resource financial project performance evaluation method and system. Background Art

[0002] In natural resource management and ecological restoration projects, performance evaluation is a key link that can effectively measure whether the project has achieved the predetermined ecological restoration goals and help project managers track and adjust the restoration progress and effects. Performance evaluation can not only provide a basis for project optimization, but also provide valuable experience for the implementation of similar projects in the future.

[0003] However, traditional ecological restoration performance evaluation methods usually rely on manual surveys and on-site surveys, facing a series of problems that seriously affect the efficiency and accuracy of the evaluation. First, manual surveys are inefficient. In large-scale ecological restoration projects, the restoration area usually covers a wide geographical range, and investigators need to visit the restoration areas one by one, measure the area, record the data manually, and conduct step-by-step statistics and analysis. This process is very time-consuming, resulting in low overall work efficiency. Secondly, manual surveys often rely on investigators to record on-site and fill in data forms. This process is easily affected by many factors, resulting in the inability to guarantee the accuracy of the data and inaccurate performance evaluation. Summary of the invention

[0004] The purpose of the present invention is to solve the problems of low efficiency and low accuracy mentioned in the above background technology, and to propose a natural resource financial project performance evaluation method and system.

[0005] The first aspect of the present invention provides a method for evaluating the performance of natural resource financial projects, the method comprising: Determine the ecological restoration indicators of the target project based on the preset performance goals; the ecological restoration indicators include cultivated land area indicators and forest area indicators; According to the preset project plan, the ecological restoration indicators are decomposed to obtain the process goals at different stages; Based on high-altitude and low-altitude remote sensing, the project construction process is monitored to obtain the ecological restoration area at the current stage; the ecological restoration area includes the newly added forest area and the newly added cultivated land area; Compare the ecological restoration area in the current stage with the process goals and generate a performance report for the current stage.

[0006] Optionally, the project construction process is monitored based on high-altitude remote sensing and low-altitude remote sensing to obtain the ecological restoration area at the current stage, including: Acquire the high-altitude remote sensing image of the target area in the current stage as the stage image; Compare the stage image with a reference image to obtain a changed area; the reference image is a high-altitude remote sensing image of the target area before the project construction; According to the change areas of the previous stage and the current stage, the stage change area is obtained; Acquire low-altitude remote sensing images of the phase-changing area as images to be classified; Identify the image to be classified to obtain the land cover types and newly added areas of different regions; the land cover types include cultivated land, forest land and others; Based on the newly added areas of cultivated land and forest land, the ecological restoration area in the current stage is obtained.

[0007] Optionally, comparing the stage image with the reference image to obtain the changed area includes: Preprocessing the stage image to obtain a stage image that is consistent with the reference image space; The processed stage image and the reference image are used as the input of the pre-trained change detection model to obtain a binary mask image and determine the image coordinates of the changed area; The image coordinates of the change area are mapped to geographic coordinates to form a geographic polygon of the change area.

[0008] Optionally, the low-altitude remote sensing image of the area of ​​phase change is acquired as the image to be classified and includes: Plan the flight path and collection interval of the drone based on the geographic information of the phase-changing area; Use drones to take aerial photos based on the flight path and collection interval to obtain multiple low-altitude remote sensing images; According to the sensor parameters and flight data of the UAV, each low-altitude remote sensing image is geometrically calibrated to obtain an orthophoto image set; Multiple images are stitched together according to the geographic information of the orthophoto images to obtain the image to be classified.

[0009] Optionally, the identifying the image to be classified to obtain the land cover types of different regions and their newly added areas includes: According to a preset size, the image to be classified is segmented to obtain a plurality of window images; The target window image is used as the input of the pre-trained classification detection model to obtain a classification result map of the target window image; the target window image is any window image; each pixel value of the classification result map is a type code corresponding to the pixel; splicing the classification result images of the multiple window images to obtain a first result image; According to each pixel value in the first result image, a target type pixel set is extracted; the target type is forest land or cultivated land; Calculate the distance between target type pixels, connect two target type pixels whose distance is less than a preset threshold, and obtain a closed area set; According to the closed area set, the newly added area of ​​the target type is calculated.

[0010] A second aspect of the present invention provides a natural resource financial project performance evaluation system, the system comprising: An indicator determination module is used to determine the ecological restoration indicators of the target project according to the preset performance goals; the ecological restoration indicators include cultivated land area indicators and forest area indicators; An indicator decomposition module is used to decompose the ecological restoration indicators according to a preset project plan to obtain process goals at different stages; The monitoring module is used to monitor the project construction process based on high-altitude remote sensing and low-altitude remote sensing to obtain the ecological restoration area at the current stage; the ecological restoration area includes the newly added forest area and the newly added cultivated land area; The data analysis module is used to compare the ecological restoration area in the current stage with the process target and generate a performance report for the current stage.

[0011] Optionally, the monitoring module includes: The first data acquisition module is used to acquire a high-altitude remote sensing image of the target area in the current stage as a stage image; A change detection module is used to compare the stage image with a reference image to obtain a changed area; the reference image is a high-altitude remote sensing image of the target area before the project construction; A new change module is added to obtain the stage change area based on the change areas of the previous stage and the current stage; The second data acquisition module is used to acquire low-altitude remote sensing images of the phase change area as images to be classified; A classification calculation module is used to identify the image to be classified to obtain the land cover types and newly added areas of different regions; the land cover types include cultivated land, forest land and others; The area accumulation module is used to obtain the ecological restoration area in the current stage based on the newly added areas of cultivated land and forest land.

[0012] Optionally, the change detection module includes: A registration module, used for preprocessing the stage image to obtain a stage image that is consistent with the reference image space; An image detection module is used to use the processed stage image and the reference image as inputs of a pre-trained change detection model to obtain a binary mask image and determine the image coordinates of the changed area; The geographic mapping module is used to map the image coordinates of the change area to geographic coordinates to form a geographic polygon of the change area.

[0013] Optionally, the second data acquisition module includes: The path planning module is used to plan the flight path and collection interval of the drone according to the geographic information of the stage-changing area; The data acquisition module is used to perform drone aerial photography according to the flight path and acquisition interval to obtain multiple low-altitude remote sensing images; The calibration and alignment module is used to perform geometric calibration on each low-altitude remote sensing image according to the sensor parameters and flight data of the UAV to obtain an orthophoto image set; The stitching module is used to stitch multiple images according to the geographic information of the orthophoto image to obtain the image to be classified.

[0014] Optionally, the classification calculation module includes: An image segmentation module is used to segment the image to be classified according to a preset size to obtain a plurality of window images; A classification recognition module is used to take the target window image as the input of the pre-trained classification detection model to obtain a classification result map of the target window image; the target window image is any window image; each pixel value of the classification result map is a type code corresponding to the pixel; A result synthesis module, used for splicing the classification result images of multiple window images to obtain a first result image; A pixel integration module is used to extract a target type pixel set according to each pixel value in the first result image; the target type is forest land or cultivated land; The region classification module is used to calculate the distance between target type pixels and connect two target type pixels whose distance is less than a preset threshold to obtain a closed region set; The area calculation module is used to calculate the newly added area of ​​the target type according to the closed area set.

[0015] Beneficial effects of the present invention: The present invention proposes a performance evaluation method for natural resource financial projects, which includes: determining ecological restoration indicators of target projects according to preset performance targets; the ecological restoration indicators include cultivated land area indicators and forest area indicators; decomposing the ecological restoration indicators according to a preset project plan to obtain process targets at different stages; monitoring the project construction process according to high-altitude remote sensing and low-altitude remote sensing to obtain the ecological restoration area at the current stage; the ecological restoration area includes newly added forest area and newly added cultivated land area; comparing the ecological restoration area at the current stage with the process target to generate a performance report for the current stage.

[0016] Monitoring through high-altitude remote sensing and low-altitude remote sensing technology can quickly obtain ecological restoration data in a wide area, avoiding the inefficient process of manual visits and greatly improving the efficiency of data collection. Quantitative analysis of remote sensing data avoids possible errors in manual data reporting, thereby ensuring the accuracy of performance evaluation. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The present invention will be further described below in conjunction with the accompanying drawings.

[0018] Figure 1 A flow chart of a method for evaluating the performance of a natural resource financial project is provided for an embodiment of the present invention; Figure 2 An architectural diagram of a natural resource financial project performance evaluation system is provided for an embodiment of the present invention. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] The embodiment of the present invention provides a method for evaluating the performance of natural resource financial projects. Figure 1 , Figure 1 A flowchart of a natural resource financial project performance evaluation method provided by an embodiment of the present invention. The method comprises the following steps: S101. Determine the ecological restoration indicators for the target project based on the preset performance goals.

[0021] S102: According to the preset project plan, the ecological restoration indicators are decomposed to obtain the process goals at different stages.

[0022] S103, monitor the project construction process based on high-altitude remote sensing and low-altitude remote sensing to obtain the ecological restoration area at the current stage.

[0023] S104, compare the ecological restoration area of ​​the current stage with the process target and generate a performance report of the current stage.

[0024] Among them, ecological restoration indicators include cultivated land area indicators and forest area indicators; the ecological restoration area includes newly added forest land area and newly added cultivated land area.

[0025] Based on a natural resource financial project performance evaluation method provided by an embodiment of the present invention, high-altitude remote sensing and low-altitude remote sensing technology are used for monitoring, and ecological restoration data in a wide area can be quickly obtained, avoiding the inefficient process of manual access, and greatly improving the efficiency of data collection. Quantitative analysis of remote sensing data avoids possible errors in manual data reporting, thereby ensuring the accuracy of performance evaluation.

[0026] In one implementation method, by setting clear performance goals and breaking them down into process goals at different stages, it can be ensured that each link of the project implementation has clear standards and evaluation basis. At the same time, the use of high-altitude remote sensing and low-altitude remote sensing technology for monitoring can provide accurate data support, improve the scientificity and accuracy of project performance evaluation, and promptly detect deviations from the predetermined goals and make adjustments.

[0027] In one implementation method, by comparing and generating reports at different stages, we can ensure that the project can obtain timely feedback during the implementation process, thereby optimizing the management and implementation effects. In addition, the performance report of each stage can clearly reflect the progress of the project, increase the transparency of project implementation, help the government and other relevant parties to supervise the progress of the project, and strengthen the accountability mechanism. In one embodiment, step S103 includes: Step 1: Obtain a high-altitude remote sensing image of the target area in the current stage as the stage image.

[0028] Step 2: Compare the stage image with the reference image to obtain the changed area.

[0029] Step three, obtain the stage change area according to the change area of ​​the previous stage and the current stage.

[0030] Step 4: Obtain low-altitude remote sensing images of the phase change area as images to be classified.

[0031] Step 5: Identify the images to be classified and obtain the land cover types and newly added areas of different regions.

[0032] Step six: Obtain the ecological restoration area at the current stage based on the newly added areas of cultivated land and forest land.

[0033] Among them, the baseline image is the high-altitude remote sensing image of the target area before the project construction; the surface cover types include cultivated land, forest land and others.

[0034] In one implementation, remote sensing technology can cover a wider area and reduce labor costs and time consumption compared to traditional ground survey methods. High-altitude remote sensing can use satellite remote sensing. By obtaining high-altitude remote sensing images and comparing them with baseline images, it can clearly identify the areas of change at different stages, provide a comprehensive understanding of changes in the project area, and effectively reduce the exploration area of ​​drones, thereby reducing energy consumption and computing resources.

[0035] In one implementation, the introduction of low-altitude remote sensing images allows for more refined identification and classification of areas of change, making it possible to clearly distinguish between cultivated land, forest land, and other land cover types, and calculate the new areas. By quantitatively measuring the new areas of cultivated land and forest land, the progress of the project can be effectively assessed.

[0036] In one embodiment, the stage image is compared with the reference image, and the changed area includes: Step 1: preprocess the stage image to obtain a stage image that is consistent with the reference image space.

[0037] Step 2: The processed stage image and the reference image are used as inputs of the pre-trained change detection model to obtain a binary mask image and determine the image coordinates of the changed area.

[0038] Step three, map the image coordinates of the change area to geographic coordinates to form a geographic polygon of the change area.

[0039] In one implementation, the accuracy of image comparison is ensured by preprocessing the stage image to make it consistent with the reference image in space. This step eliminates errors that may be caused by differences in image resolution, angle or time, and provides a reliable data basis for subsequent change detection.

[0040] In one implementation, the change detection model may be a convolutional neural network (CNN) model. By using a pre-trained change detection model to compare images, the changed area can be automatically identified. This method not only improves detection efficiency, but also reduces the error and complexity of manual operation, and can quickly and accurately obtain the location and range of the changed area.

[0041] In one embodiment, obtaining low-altitude remote sensing images of the phase change region as images to be classified includes: Step 1: Plan the flight path and collection interval of the drone based on the geographic information of the phase-change area.

[0042] Step 2: Use a drone to take aerial photos based on the flight path and collection interval to obtain multiple low-altitude remote sensing images.

[0043] Step three: According to the sensor parameters and flight data of the UAV, each low-altitude remote sensing image is geometrically calibrated to obtain an orthophoto image set.

[0044] Step 4: stitch multiple images according to the geographic information of the orthophoto image to obtain the image to be classified.

[0045] In one implementation, the flight path of the drone should cover the entire change area with redundancy at the edges. The acquisition interval should meet the minimum overlap rate between images, such as 30%, to ensure that the resulting image set can cover the entire change area.

[0046] In one implementation, drone aerial photography can provide higher resolution images and greater detail capture capabilities than satellite remote sensing, thereby improving the precision and accuracy of classification.

[0047] In one implementation, overlapping areas can be eliminated by stitching, thus avoiding repeated calculations caused by classifying a single image.

[0048] In one embodiment, the image to be classified is identified to obtain the land cover types and newly added areas of different regions, including: Step 1: Segment the image to be classified according to a preset size to obtain multiple window images.

[0049] Step 2: Use the target window image as the input of the pre-trained classification detection model to obtain a classification result map of the target window image.

[0050] Step three, splicing the classification result images of multiple window images to obtain a first result image.

[0051] Step 4: extract a target type pixel set according to each pixel value in the first result image.

[0052] Step 5: Calculate the distance between target type pixels, connect two target type pixels whose distance is less than a preset threshold, and obtain a closed area set.

[0053] Step 6: Calculate the newly added area of ​​the target type based on the closed area set.

[0054] Among them, the target window image is any window image; each pixel value of the classification result image is the type code corresponding to the pixel; the target type is forest land or cultivated land.

[0055] In one implementation, the classification detection model may be a UNet model. The UNet model is a classic deep learning structure that is particularly suitable for image segmentation tasks. The model can accurately segment the image to be classified according to different surface cover types.

[0056] In one implementation, the classification result images of multiple window images are spliced ​​to obtain a first result image, which can ensure the consistency of classification information of different areas in the image. The spliced ​​result image provides a complete spatial view, which can more clearly reflect the different surface cover types and their distribution in the entire image.

[0057] In one implementation, by calculating the distance between target type pixels, connecting pixels whose distance is less than a preset threshold to obtain a closed area set, this method can effectively identify connected areas in the image. This can clearly define each type of continuous area, which helps to perform accurate area calculation.

[0058] The embodiment of the present invention provides a natural resource financial project performance evaluation system. Figure 2 , Figure 2 This is a diagram of the architecture of a natural resource financial project performance evaluation system provided by an embodiment of the present invention. The system includes: The indicator determination module is used to determine the ecological restoration indicators of the target project based on the preset performance goals.

[0059] The indicator decomposition module is used to decompose the ecological restoration indicators according to the preset project plan to obtain the process goals at different stages.

[0060] The monitoring module is used to monitor the project construction process based on high-altitude remote sensing and low-altitude remote sensing to obtain the ecological restoration area at the current stage.

[0061] The data analysis module is used to compare the ecological restoration area in the current stage with the process target and generate a performance report for the current stage.

[0062] Among them, ecological restoration indicators include cultivated land area indicators and forest area indicators; the ecological restoration area includes newly added forest land area and newly added cultivated land area.

[0063] Based on the natural resource financial project performance evaluation system provided by the embodiment of the present invention, through high-altitude remote sensing and low-altitude remote sensing technology for monitoring, it is possible to quickly obtain ecological restoration data in a wide area, avoiding the inefficient process of manual access, and greatly improving the efficiency of data collection. Through quantitative analysis of remote sensing data, it avoids possible errors in manual data reporting, thereby ensuring the accuracy of performance evaluation.

[0064] In one embodiment, the monitoring module includes: The first data acquisition module is used to acquire a high-altitude remote sensing image of the target area in the current stage as a stage image; The change detection module is used to compare the stage image with the reference image to obtain the changed area; the reference image is a high-altitude remote sensing image of the target area before the project construction.

[0065] A new change module is added to obtain the stage change area based on the change areas of the previous stage and the current stage.

[0066] The second data acquisition module is used to acquire low-altitude remote sensing images of the phase change area as images to be classified.

[0067] The classification calculation module is used to identify the images to be classified and obtain the land cover types and newly added areas of different regions; the land cover types include cultivated land, forest land and others.

[0068] The area accumulation module is used to obtain the ecological restoration area in the current stage based on the newly added areas of cultivated land and forest land.

[0069] In one embodiment, the change detection module includes: The registration module is used to preprocess the stage image to obtain a stage image that is consistent with the reference image space.

[0070] The image detection module is used to use the processed stage image and the reference image as the input of the pre-trained change detection model to obtain a binary mask image and determine the image coordinates of the changed area.

[0071] The geographic mapping module is used to map the image coordinates of the change area to geographic coordinates to form a geographic polygon of the change area.

[0072] In one embodiment, the second data acquisition module includes: The path planning module is used to plan the flight path and collection interval of the drone according to the geographic information of the stage-changing area.

[0073] The data acquisition module is used to perform drone aerial photography according to the flight path and acquisition interval to obtain multiple low-altitude remote sensing images.

[0074] The calibration and alignment module is used to perform geometric calibration on each low-altitude remote sensing image according to the sensor parameters and flight data of the UAV to obtain an orthophoto image set.

[0075] The stitching module is used to stitch multiple images according to the geographic information of the orthophoto image to obtain the image to be classified.

[0076] In one embodiment, the classification calculation module includes: The image segmentation module is used to segment the image to be classified according to a preset size to obtain multiple window images.

[0077] The classification recognition module is used to take the target window image as the input of the pre-trained classification detection model to obtain a classification result map of the target window image; the target window image is any window image; each pixel value of the classification result map is the type code corresponding to the pixel.

[0078] The result synthesis module is used to splice the classification result images of multiple window images to obtain a first result image.

[0079] The pixel integration module is used to extract a target type pixel set according to each pixel value in the first result image; the target type is forest land or cultivated land.

[0080] The region classification module is used to calculate the distance between target type pixels and connect two target type pixels whose distance is less than a preset threshold to obtain a closed region set.

[0081] The area calculation module is used to calculate the newly added area of ​​the target type based on the closed area set.

[0082] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A method for evaluating the performance of natural resource financial projects, characterized in that: The method comprises: Determine the ecological restoration indicators of the target project based on the preset performance goals; the ecological restoration indicators include cultivated land area indicators and forest area indicators; According to the preset project plan, the ecological restoration indicators are decomposed to obtain the process goals at different stages; Based on high-altitude and low-altitude remote sensing, the project construction process is monitored to obtain the ecological restoration area at the current stage; the ecological restoration area includes the newly added forest area and the newly added cultivated land area; Compare the ecological restoration area in the current stage with the process goals and generate a performance report for the current stage.

2. A natural resource financial project performance evaluation method according to claim 1, characterized in that: The project construction process is monitored based on high-altitude and low-altitude remote sensing, and the ecological restoration area at the current stage includes: Acquire the high-altitude remote sensing image of the target area in the current stage as the stage image; Compare the stage image with a reference image to obtain a changed area; the reference image is a high-altitude remote sensing image of the target area before the project construction; According to the change areas of the previous stage and the current stage, the stage change area is obtained; Acquire low-altitude remote sensing images of the phase-changing area as images to be classified; Identify the image to be classified to obtain the land cover types and newly added areas of different regions; the land cover types include cultivated land, forest land and others; Based on the newly added areas of cultivated land and forest land, the ecological restoration area in the current stage is obtained.

3. A natural resource financial project performance evaluation method according to claim 2, characterized in that: The step of comparing the stage image with the reference image to obtain the changed area comprises: Preprocessing the stage image to obtain a stage image that is consistent with the reference image space; The processed stage image and the reference image are used as the input of the pre-trained change detection model to obtain a binary mask image and determine the image coordinates of the changed area; The image coordinates of the change area are mapped to geographic coordinates to form a geographic polygon of the change area.

4. A natural resource financial project performance evaluation method according to claim 2, characterized in that: The low-altitude remote sensing images of the area where the phase changes are acquired, as images to be classified, include: Plan the flight path and collection interval of the drone based on the geographic information of the phase-changing area; Use drones to take aerial photos based on the flight path and collection interval to obtain multiple low-altitude remote sensing images; According to the sensor parameters and flight data of the UAV, each low-altitude remote sensing image is geometrically calibrated to obtain an orthophoto image set; Multiple images are stitched together according to the geographic information of the orthophoto images to obtain the image to be classified.

5. A natural resource financial project performance evaluation method according to claim 2, characterized in that: The identifying of the image to be classified to obtain the land cover types and newly added areas of different regions includes: According to a preset size, the image to be classified is segmented to obtain a plurality of window images; The target window image is used as the input of the pre-trained classification detection model to obtain a classification result map of the target window image; the target window image is any window image; each pixel value of the classification result map is a type code corresponding to the pixel; splicing the classification result images of the multiple window images to obtain a first result image; According to each pixel value in the first result image, a target type pixel set is extracted; the target type is forest land or cultivated land; Calculate the distance between target type pixels, connect two target type pixels whose distance is less than a preset threshold, and obtain a closed area set; According to the closed area set, the newly added area of ​​the target type is calculated.

6. A natural resource financial project performance evaluation system, characterized in that: The system comprises: An indicator determination module is used to determine the ecological restoration indicators of the target project according to the preset performance goals; the ecological restoration indicators include cultivated land area indicators and forest area indicators; An indicator decomposition module is used to decompose the ecological restoration indicators according to a preset project plan to obtain process goals at different stages; The monitoring module is used to monitor the project construction process based on high-altitude remote sensing and low-altitude remote sensing to obtain the ecological restoration area at the current stage; the ecological restoration area includes the newly added forest area and the newly added cultivated land area; The data analysis module is used to compare the ecological restoration area in the current stage with the process target and generate a performance report for the current stage.

7. A natural resource financial project performance evaluation system according to claim 6, characterized in that: The monitoring module comprises: The first data acquisition module is used to acquire a high-altitude remote sensing image of the target area in the current stage as a stage image; A change detection module is used to compare the stage image with a reference image to obtain a changed area; the reference image is a high-altitude remote sensing image of the target area before the project construction; A new change module is added to obtain the stage change area based on the change areas of the previous stage and the current stage; The second data acquisition module is used to acquire low-altitude remote sensing images of the phase change area as images to be classified; A classification calculation module is used to identify the image to be classified to obtain the land cover types and newly added areas of different regions; the land cover types include cultivated land, forest land and others; The area accumulation module is used to obtain the ecological restoration area in the current stage based on the newly added areas of cultivated land and forest land.

8. A natural resource financial project performance evaluation system according to claim 7, characterized in that: The change detection module comprises: A registration module, used for preprocessing the stage image to obtain a stage image that is consistent with the reference image space; An image detection module is used to use the processed stage image and the reference image as inputs of a pre-trained change detection model to obtain a binary mask image and determine the image coordinates of the changed area; The geographic mapping module is used to map the image coordinates of the change area to geographic coordinates to form a geographic polygon of the change area.

9. A natural resource financial project performance evaluation system according to claim 7, characterized in that: The second data acquisition module includes: The path planning module is used to plan the flight path and collection interval of the drone according to the geographic information of the stage-changing area; The data acquisition module is used to perform drone aerial photography according to the flight path and acquisition interval to obtain multiple low-altitude remote sensing images; The calibration and alignment module is used to perform geometric calibration on each low-altitude remote sensing image according to the sensor parameters and flight data of the UAV to obtain an orthophoto image set; The stitching module is used to stitch multiple images according to the geographic information of the orthophoto image to obtain the image to be classified.

10. A natural resource financial project performance evaluation system according to claim 7, characterized in that: The classification calculation module includes: An image segmentation module is used to segment the image to be classified according to a preset size to obtain a plurality of window images; A classification recognition module is used to take the target window image as the input of the pre-trained classification detection model to obtain a classification result map of the target window image; the target window image is any window image; each pixel value of the classification result map is a type code corresponding to the pixel; A result synthesis module, used for splicing the classification result images of multiple window images to obtain a first result image; A pixel integration module is used to extract a target type pixel set according to each pixel value in the first result image; the target type is forest land or cultivated land; The region classification module is used to calculate the distance between target type pixels and connect two target type pixels whose distance is less than a preset threshold to obtain a closed region set; The area calculation module is used to calculate the newly added area of ​​the target type according to the closed area set.

Images