Ecological restoration area remote sensing identification method and device

Abstract

The invention belongs to the technical field of remote sensing vegetation detection, and provides an ecological restoration area remote sensing identification method and device, and the method comprises the steps: obtaining remote sensing data and actual measurement data of a target area; aiming at remote sensing data, acquiring NDVRI index images at different time in a historical stage; the NDVRI index is a normalized sandy land-vegetation restoration index, when the NDVRI index is increased, it is indicated that the bare soil or sandy land range is expanded, and when the NDVRI index is decreased, it is indicated that vegetation coverage is increased; constructing a prediction model by using the NDVRI images at different time in the historical stage, and predicting the NDVRI at the moment t in the monitoring stage; and screening out a breakpoint of which the variation amplitude between the predicted NDVRI index and the actually measured data is less than 0, and determining an ecological restoration area. The method is suitable for vegetation monitoring of the sand land environment, the comprehensive characteristics of vegetation recovery and bare land reduction are emphasized by integrating the brightness and greenness components, and the precision and reliability of sand land ecological monitoring are improved.

Description

Technical Field

[0001] The invention relates to the technical field of remote sensing vegetation monitoring, and in particular to a remote sensing identification method and device for an ecological restoration area. Background Art

[0002] Desertification, as one of the prominent environmental problems in grassland areas, has become an ecological crisis that needs to be solved urgently. In order to curb the harm of wind and sand and protect the ecological environment of sandy land, ecological restoration projects with vegetation reconstruction as the core have been widely used. These projects have played an important role in alleviating regional desertification and promoting sandy ecological restoration. At the same time, establishing a scientific and effective ecological restoration monitoring method is of vital significance for evaluating the effectiveness of restoration and optimizing subsequent management strategies.

[0003] Traditional vegetation change monitoring methods mainly include ground measurement and remote sensing monitoring. Although ground measurement methods such as visual estimation, sampling and photography have high accuracy, they are difficult to obtain comprehensive information on sand restoration at a regional scale due to limitations in manpower and material resources. In contrast, remote sensing monitoring methods have gradually become the main means of monitoring ecological restoration due to their wide coverage and high temporal resolution. In remote sensing time series data, surface changes are usually affected by seasonal changes, trend changes and sudden changes, and are also restricted by interference factors such as atmospheric scattering, cloud shadow effects and geometric errors. Therefore, how to effectively determine and identify surface changes in complex remote sensing time series data has become a key issue in remote sensing time series analysis.

[0004] In recent years, remote sensing time series analysis methods have developed rapidly, and a variety of technical paths have been formed, including threshold method, difference method, trajectory segmentation method, classification method, statistical method and regression method. These methods have their own advantages in the characteristics of time series data frequency, spectral index type, detection unit and real-time performance, but they also have certain limitations. For example, trajectory segmentation algorithms (such as LandTrendr) mainly use the univariate spectral index NBR to detect surface changes, but they rely on annual mean data, making it difficult to capture continuous changes, and the real-time detection performance is poor; threshold algorithms (such as VCT) identify changes through the univariate spectral index IFZ, which is also limited by the limitations of low time series frequency; and the CCDC algorithm, which combines statistics and classification, uses multivariate spectral input and can detect changes in land cover types in near real time, but multivariate input is prone to introduce errors, and the complexity of time series data affects its detection accuracy. Summary of the invention

[0005] In view of the defects in the prior art, the present invention provides a remote sensing identification method and device for ecological restoration areas to solve the problem that the accuracy of vegetation cover change detection in sandy areas is affected by the temporal limitations of index performance and the index data set used, resulting in limitations.

[0006] In a first aspect, the present invention provides a method for remote sensing identification of ecological restoration areas, comprising: Obtain remote sensing data and measured data of the target area; For the remote sensing data, NDVRI index images at different times in the historical stage are obtained; the NDVRI index is a normalized sand-vegetation restoration index. When the NDVRI index increases, it indicates that the scope of bare soil or sand is expanding, and when the NDVRI index decreases, it indicates that the vegetation coverage is increasing; Using the NDVRI index images at different times in the historical stage, a prediction model is constructed to predict the NDVRI index at time t in the monitoring stage; The breakpoints where the variation range of the NDVRI index is less than 0 are screened out to determine the ecological restoration area.

[0007] It can be seen from the above technical scheme that the present invention provides a remote sensing identification method for ecological restoration areas, and proposes a normalized sand-vegetation restoration index combined with tasseled cap transformation. This index captures the dynamic changes of vegetation recovery and ground reduction during sand restoration by integrating brightness and greenness components. It can effectively identify the temporal and spatial characteristics of dynamic changes in restoration, and provides a novel and efficient technical means for the scientific evaluation and dynamic monitoring of sand ecological restoration.

[0008] Optionally, acquiring remote sensing data of the target area includes: Collect all available Sentinel-2 data in the target area in the B1-B12 and B8A bands; Splice multiple images acquired at the same time into a remote sensing image that completely covers the target area; Converting the remote sensing data value of the remote sensing image into a reflectivity value; Based on the model, the atmospheric scattering and absorption effects are removed to obtain the surface reflectivity value.

[0009] Optionally, the NDVRI index is determined based on the surface reflectance value after tasseled cap transformation, and is specifically calculated according to the following formula: , where B is the brightness component of the surface reflectance value after tasseled cap transformation, and G is the greenness component of the surface reflectance value after tasseled cap transformation.

[0010] Optionally, the use of NDVRI index images at different times in the historical stage to construct a prediction model to predict the NDVRI index at time t in the monitoring stage includes: A regression model is constructed based on historical data; the regression model is used to obtain the predicted value of the NDVRI index at time t ; Get the moving residual and , determine whether a breakpoint occurs; When the breakpoint is determined, the actual observed value at the time of the breakpoint and the predicted value of the NDVRI index are calculated. The difference between them determines the magnitude of change; Based on the magnitude of the changes, determine the ecological restoration area.

[0011] Optionally, the regression model is constructed using ordinary least squares method, and the predicted value of NDVRI index at time t is Specifically , in, is the intercept coefficient estimated based on historical period observations, represents the long-term trend term, k represents the harmonic term order, is the amplitude, To harmonize the seasons, is the frequency, the sine function describes the seasonal fluctuations, is the random error at time t.

[0012] Optionally, the moving residual and The MOSUM method is used to determine the specific calculation according to the following formula: , in, is the estimated value of the variance, n is the number of observations in the historical stage, h is the size of the sliding window, h and n are fixed ratios, and t represents the time point.

[0013] Optionally, the step of screening out the breakpoints where the variation between the predicted NDVRI index and the measured data is less than 0 and determining the ecological restoration area is specifically as follows: The change range is based on the predicted value of the NDVRI index at time t and the remote sensing measured value at time t Determine that the variation range is ; When the change amplitude mc is less than 0, it is determined as an ecological restoration area.

[0014] In a second aspect, the present invention provides an ecological restoration area remote sensing identification device, comprising: Acquisition module, used to obtain remote sensing data and measured data of the target area; An index processing module is used to obtain NDVRI index images at different times in the historical stage for the remote sensing data; the NDVRI index is a normalized sand-vegetation restoration index. When the NDVRI index increases, it indicates that the scope of bare soil or sand is expanding, and when the NDVRI index decreases, it indicates that the vegetation coverage is increasing; A prediction module is used to construct a prediction model using NDVRI index images at different times in the historical stage to predict the NDVRI index at time t in the monitoring stage; The restoration determination module is used to screen out the breakpoints where the change between the predicted NDVRI index and the measured data is less than 0, and determine the ecological restoration area.

[0015] In a third aspect, an embodiment of the present invention provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of any of the above methods when executing the computer program.

[0016] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having computer program instructions stored thereon, which implement the steps of any of the above methods when executed by a processor.

[0017] By adopting the above technical solution, the present application has the following beneficial effects: The present invention proposes a normalized sand-vegetation restoration index NDVRI based on tasseled cap transform, and combines it with a prediction model to dynamically monitor the time series changes of sand restoration. The NDVRI index emphasizes the comprehensive characteristics of vegetation restoration and bare land reduction by integrating brightness (B) and greenness (G) components, and shows significant advantages in spatial accuracy and temporal response characteristics, reducing the impact of the high reflectivity of the sand background on the detection results. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following is a brief introduction to the drawings required for the specific embodiments or the description of the prior art. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn according to the actual scale.

[0019] Figure 1 A flowchart of a remote sensing identification method for an ecological restoration area provided by an embodiment of the present invention is shown; Figure 2 A schematic diagram showing an ecological restoration area identified by the NDVRI index provided by an embodiment of the present invention is shown; Figure 3 A schematic diagram showing the ecological restoration time identified by the NDVRI index provided in an embodiment of the present invention is shown; Figure 4 A schematic diagram showing the average hysteresis time of the NDVRI index provided by an embodiment of the present invention is shown; Figure 5 A schematic diagram of a remote sensing identification device for an ecological restoration area provided by an embodiment of the present invention is shown; Figure 6 A schematic diagram of an electronic device provided by an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0020] The following embodiments of the technical solution of the present invention are described in detail in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and are therefore only used as examples, and cannot be used to limit the protection scope of the present invention. It should be noted that, unless otherwise specified, the technical terms or scientific terms used in this application should have the common meanings understood by those skilled in the art to which the present invention belongs.

[0021] At present, in the process of capturing sand restoration, due to the limited ability of various indices to describe low-coverage green vegetation information in arid and semi-arid areas, in the process of evaluating and monitoring vegetation restoration in sand, it is easily affected by the background of bare land and it is difficult to accurately reflect the actual changes. Based on this, a vegetation change detection method for sand environment is proposed to make up for the shortcomings of traditional methods and improve the accuracy and reliability of sand ecological monitoring.

[0022] like Figure 1 As shown, this embodiment provides a remote sensing identification method for ecological restoration areas, including: S1. Obtain remote sensing data and measured data of the target area.

[0023] S2. Obtain NDVRI index images at different times in the historical stage for remote sensing data; the NDVRI index is the normalized sand-vegetation restoration index. When the NDVRI index rises, it indicates that the scope of bare soil or sand is expanding, and when the NDVRI index decreases, it indicates that the vegetation coverage is increasing.

[0024] In this step, during the dataset construction process, the NDVRI index value is first calculated for each image, and the pixel median method is used to synthesize the monthly NDVRI index image. The historical stage is the stage before the ecological restoration project is carried out, and the monitoring stage is the stage after the ecological restoration project has begun.

[0025] At the same time, in order to highlight the superiority of the NDVRI index, a time series data set of the traditional vegetation index NDVI and SAVI was constructed for comparison. The formula is as follows: , Among them, NIR and RED represent the reflectance of near-infrared and red bands, respectively.

[0026] The SAVI (Soil Adjustment Index) formula is: , Where L is the adjustment factor (usually 0.5).

[0027] Before predicting the NDVRI index based on the forecast model, in order to further improve the quality and continuity of the time series data, the Savitzky-Golay filter (SG filter) is also used to smooth the synthetic monthly time series data.

[0028] S3. Use the NDVRI index images at different times in the historical stage to build a prediction model and predict the NDVRI index at time t in the monitoring stage.

[0029] S4. Filter out the breakpoints where the change range of NDVRI index is less than 0 and determine the ecological restoration area.

[0030] Optionally, obtaining remote sensing data of the target area in S1 includes: S101. Collect all available Sentinel-2 data in the target area in the B1-B12 and B8A bands.

[0031] S102. Stitching multiple images acquired at the same time into a remote sensing image that completely covers the target area.

[0032] S103. Convert the remote sensing data value of the remote sensing image into a reflectivity value.

[0033] This step is used to convert the value obtained by the sensor into the reflectivity value or the radiation brightness value required in this embodiment. Generally, the DN value obtained by remote sensing technology is converted in this step.

[0034] S104. Remove the effects of atmospheric scattering and absorption based on a model (such as 6S or FLAASH) to obtain the surface reflectance value.

[0035] The data in step S1 include remote sensing data and measured data. The remote sensing data is used to collect the B1-B12 and B8A bands of all available sentinel data in the target area. The measured data is the location and time of the area that has undergone ecological restoration.

[0036] Optionally, the NDVRI index is determined based on the surface reflectance value after tasseled cap transformation, and is specifically calculated according to the following formula: , where B is the brightness component of the surface reflectance value after the tasseled cap transformation, and G is the greenness component of the surface reflectance value after the tasseled cap transformation.

[0037] The NDVRI index significantly enhances the contrast signal between bare soil and vegetation through the difference between brightness and green components, thereby improving the ability to identify dynamic changes in sand. The index uses a ratio form, and through the weighting of the denominator, it balances the numerical differences in brightness and greenness, effectively reducing fluctuations caused by environmental factors such as imaging angles and atmospheric conditions. During the sand restoration process, the dynamic changes in brightness and green components reflect the process of transformation from bare soil to vegetation coverage. Specifically, when the NDVRI index rises, it usually indicates that the scope of bare soil or sand has expanded and the surface brightness component has increased; when the index decreases, it indicates that vegetation coverage has increased and the green component is dominant.

[0038] Specifically, the surface reflectance value is transformed by the tasseled cap to obtain the brightness component and greenness component of the NDVRI index calculated in this step. Tasseled Cap Transformation (TCT), also known as KT Transformation, is a classic linear transformation method used to compress and reconstruct the multispectral pixel values ​​of remote sensing images. Since the tasseled cap transformation coefficients depend on the band characteristics of the sensor, its parameters are usually not directly applicable to different sensors. To ensure the reliability of the results, this embodiment refers to a variety of coefficient derivation methods, and finally adopts the Sentinel-2 tasseled cap transformation coefficients derived by Roumen Nedkov et al. Generally, the main components extracted by the tasseled cap transformation include brightness (Brightness, TCB), greenness (Greenness, TCG) and wetness (Wetness, TCW), which respectively characterize the surface reflectance characteristics, vegetation coverage and soil moisture level. In this embodiment, only the greenness component and the brightness component are used. The coefficients used to calculate each component are shown in Table 1.

[0039] Table 1 Coefficients used for sentinel 2 tasseled cap transformation

[0040] Table 1 (continued)

[0041] Optionally, step S3 includes: S301. Construct a regression model based on historical data; the regression model is used to obtain the predicted value of the NDVRI index at time t .

[0042] Specifically, the regression model is constructed using the ordinary least squares method, and the predicted value of the NDVRI index at time t is Specifically , in, is the intercept coefficient estimated based on historical period observations, represents the long-term trend term, k represents the harmonic term order, is the amplitude, To harmonize the seasons, is the frequency, the sine function describes the seasonal fluctuations, is the random error at time t. The regression model includes trend terms and seasonal terms to adapt to the diversity of time series.

[0043] S302. Obtain the moving residual and , determine whether a breakpoint occurs; Optionally, the moving residual and The MOSUM method is used to determine the specific calculation according to the following formula: , in, is the estimated value of the variance, n is the number of observations in the historical stage, h is the size of the sliding window, h and n are fixed ratios, and t represents the time point.

[0044] when Exceeding the significance level , you can determine that a breakpoint has occurred.

[0045] S303. When a breakpoint is determined, the actual observed value at the time of the breakpoint and the predicted value of the NDVRI index are calculated. The difference between them determines the magnitude of change; S304. Determine the ecological restoration area based on the extent of change.

[0046] Optionally, the breakpoints where the variation between the predicted NDVRI index and the measured data is less than 0 are screened out to determine the ecological restoration area, specifically: The change range is based on the predicted value of the NDVRI index at time t and the remote sensing measured value at time t Determine, the change range is ; When the change amplitude mc is less than 0, it is determined as an ecological restoration area.

[0047] In the formula is the actual observation value at time t. In terms of parameter setting, in order to make the results comparable, the second-order (k=2) harmonic model is used. The bandwidth of the MOSUM detection process is set to h=0.25, and the significance level is set to α=0.05. In this embodiment, because only the ecological restoration scenario is concerned, the NDVRI only considers the case where the power outage amplitude mc is less than 0, and the SAVI and NDVI indexes only consider the case where the power outage amplitude is greater than 0.

[0048] Figure 2-3 Based on the method provided in this embodiment, a specific analysis is performed by taking Xiwuzhumuqin Banner as an example. Figure 2 The ecological restoration area identified by the NDVRI index. Figure 3 The time of ecological restoration identified by the NDVRI index. Specifically, the historical stage is set as 2016-2019, and the monitoring stage is set as 2020-2024.

[0049] The method provided in this embodiment is simulated and verified as follows.

[0050] (1) Spatial accuracy evaluation The overall accuracy (OA), intersection over union (IOU), recall (REC), precision (PRE), and F1 score (F1) are considered horizontally as evaluation indicators to verify the performance effect of the method provided in this embodiment.

[0051] In the change detection task, a higher precision indicates a lower error rate, and a higher recall indicates a lower missed detection rate. OA, IOU, and F1 are comprehensive indicators of the performance of the reaction index. The higher the corresponding value, the better the performance of the index. The specific description of each evaluation index is as follows:

[0052]

[0053]

[0054]

[0055]

[0056] Among them, TP is the number of correctly classified restored pixels, TN is the number of correctly classified unrestored pixels, FP is the number of incorrectly classified restored pixels, and FN is the number of incorrectly classified unrestored pixels. The spatial accuracy results are shown in Table 2: Table 2 Comparison results of spatial accuracy of different indexes

[0057] (2) Time accuracy evaluation In order to quantify and compare the lag time of different remote sensing indices in the process of ecological restoration, we used the lag time curve method. Specifically, the change years of ecological restoration areas were summarized using multiple vegetation indices (NDVRI, SAVI and NDVI), and the lag time of each pixel was calculated. The lag time is defined as the time difference between the actual restoration year and the year when the index change was detected. The formula is as follows:

[0058] in, is the year of change detected using remote sensing indices, It is the actual year of ecological restoration.

[0059] At the same time, in order to further quantify the time response characteristics of different indexes, the average pixel lag time of each index is calculated, as shown in Table 3. Figure 4 As shown in the figure, the average lag time of the normalized sand-vegetation restoration index NDVRI is significantly lower than that of SAVI and NDVI, and the time response characteristics of NDVRI are better than those of SAVI and NDVI indices.

[0060] Table 3 Statistics of average lag time of each index

[0061] The experimental results show that the overall accuracy (86.0%) and intersection-over-union ratio (73.5%) of the NDVRI index are better than those of SAVI (84.0%, 65.4%) and NDVI (78.4%, 55.8%). In terms of temporal response characteristics, the average lag time of NDVRI is only 1.8 years, which is better than SAVI (2.3 years) and NDVI (3.1 years), showing higher sensitivity and spatial consistency to changes in the early stages of restoration.

[0062] In one embodiment, a remote sensing identification device 50 for an ecological restoration area is provided. Figure 5 As shown, including: The acquisition module 501 is used to obtain remote sensing data and measured data of the target area; The index processing module 502 is used to obtain NDVRI index images at different times in the historical stage for the remote sensing data; the NDVRI index is a normalized sand-vegetation restoration index. When the NDVRI index increases, it indicates that the scope of bare soil or sand is expanding, and when the NDVRI index decreases, it indicates that the vegetation coverage is increasing; A prediction module 503 is used to construct a prediction model using NDVRI index images at different times in the historical stage to predict the NDVRI index at time t in the monitoring stage; The restoration determination module 504 is used to screen out the breakpoints where the variation between the predicted NDVRI index and the measured data is less than 0, and determine the ecological restoration area.

[0063] The ecological restoration area remote sensing identification device 50 provided in the embodiment of the present application adopts the same inventive concept as the above-mentioned ecological restoration area remote sensing identification method, and can achieve the same beneficial effects, which will not be repeated here.

[0064] Based on the same inventive concept as the above-mentioned ecological restoration area remote sensing identification method, the embodiment of the present application also provides an electronic device 60, such as Figure 6 As shown, the electronic device 60 may include a processor 601 and a memory 602 .

[0065] The processor 601 can implement or execute the methods, steps and logic diagrams disclosed in the embodiments of the present invention. The steps of the methods disclosed in the embodiments of the present invention can be directly embodied as being executed by a hardware processor, or by a combination of hardware and software modules in the processor.

[0066] The memory 602 is a non-volatile computer-readable storage medium that can be used to store non-volatile software programs, non-volatile computer executable programs, and modules. The memory may include at least one type of storage medium, and the memory is any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory 602 in the embodiment of the present invention can also be a circuit or any other device that can realize a storage function, and is used to store program instructions and / or data.

[0067] An embodiment of the present invention provides a computer-readable storage medium for storing computer program instructions used by the above-mentioned electronic device, which includes a program for executing the above-mentioned ecological restoration area remote sensing identification method.

[0068] The above embodiments are only used to introduce the technical solutions of the present application in detail, but the description of the above embodiments is only used to help understand the methods of the embodiments of the present invention and should not be understood as limiting the embodiments of the present invention. Any changes or substitutions that can be easily thought of by technicians in this technical field should be included in the protection scope of the embodiments of the present invention.

Claims

1. A remote sensing identification method for ecological restoration areas, characterized in that: include: Obtain remote sensing data and measured data of the target area; For the remote sensing data, NDVRI index images at different times in the historical stage are obtained; the NDVRI index is a normalized sand-vegetation restoration index. When the NDVRI index increases, it indicates that the scope of bare soil or sand is expanding, and when the NDVRI index decreases, it indicates that the vegetation coverage is increasing; Using the NDVRI index images at different times in the historical stage, a prediction model is constructed to predict the NDVRI index at time t in the monitoring stage; The breakpoints where the change between the predicted NDVRI index and the measured data is less than 0 are screened out to determine the ecological restoration area.

2. The method according to claim 1, characterized in that The step of acquiring remote sensing data of the target area includes: Collect all available Sentinel-2 data in the target area in the B1-B12 and B8A bands; Splice multiple images acquired at the same time into a remote sensing image that completely covers the target area; Converting the remote sensing data value of the remote sensing image into a reflectivity value; Based on the model, the atmospheric scattering and absorption effects are removed to obtain the surface reflectivity value.

3. The method according to claim 2, characterized in that The NDVRI index is determined based on the surface reflectance value after tasseled cap transformation, and is specifically calculated according to the following formula: , where B is the brightness component of the surface reflectance value after tasseled cap transformation, and G is the greenness component of the surface reflectance value after tasseled cap transformation.

4. The method according to claim 3, characterized in that The method of constructing a prediction model using NDVRI index images at different times in the historical stage to predict the NDVRI index at time t in the monitoring stage includes: A regression model is constructed based on historical data; the regression model is used to obtain the predicted value of the NDVRI index at time t ; Get the moving residual and , determine whether a breakpoint occurs; When the breakpoint is determined, the actual observed value at the time of the breakpoint and the predicted value of the NDVRI index are calculated. The difference between them determines the magnitude of change; Based on the magnitude of the changes, determine the ecological restoration area.

5. The method according to claim 4, characterized in that The regression model is constructed using the ordinary least squares method, and the predicted value of the NDVRI index at time t Specifically ， in, is the intercept coefficient estimated based on historical period observations, represents the long-term trend term, k represents the harmonic term order, is the amplitude, To harmonize the seasons, is the frequency, the sine function describes the seasonal fluctuations, is the random error at time t.

6. The method according to claim 5, characterized in that The moving residual and The MOSUM method is used to determine the specific calculation according to the following formula: ， in, is the estimated value of the variance, n is the number of observations in the historical stage, h is the size of the sliding window, h and n are fixed ratios, and t represents the time point.

7. The method according to claim 6, characterized in that The method of screening out the breakpoints where the variation between the predicted NDVRI index and the measured data is less than 0 and determining the ecological restoration area is specifically: The change range is based on the predicted value of the NDVRI index at time t and the remote sensing measured value at time t Determine that the variation range is ; When the change amplitude mc is less than 0, it is determined as an ecological restoration area.

8. A remote sensing identification device for ecological restoration areas, characterized in that: include: Acquisition module, used to obtain remote sensing data and measured data of the target area; An index processing module is used to obtain NDVRI index images at different times in the historical stage for the remote sensing data; the NDVRI index is a normalized sand-vegetation restoration index. When the NDVRI index increases, it indicates that the scope of bare soil or sand is expanding, and when the NDVRI index decreases, it indicates that the vegetation coverage is increasing; A prediction module is used to construct a prediction model using NDVRI index images at different times in the historical stage to predict the NDVRI index at time t in the monitoring stage; The restoration determination module is used to screen out the breakpoints where the change between the predicted NDVRI index and the measured data is less than 0, and determine the ecological restoration area.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having computer program instructions stored thereon, characterized in that: When the computer program instructions are executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images