A near real-time land degradation monitoring method and system

Abstract

The application provides a kind of near real-time land degradation monitoring method and system, the method comprises: obtaining the historical and near real-time multi-source remote sensing data of target area and carrying out space-time sequence reconstruction, obtain the spectral feature time series of each pixel point in target area;Obtain the topographic features and climate characteristics of target area and fuse with spectral feature time series, obtain the multi-dimensional feature vector of each pixel point;Obtain the historical land cover classification of each pixel point in the first historical reference period, calculate the near real-time land cover classification of each pixel point;According to the multi-dimensional feature vector of each pixel point, and / or historical land cover classification and near real-time land cover classification, calculate the land degradation evaluation index of each pixel point, and evaluate the land degradation of pixel point. Realize the quick prediction and use of near real-time data land cover classification, ensure the real-time of monitoring result, and detect land degradation from multiple aspects, form a complete monitoring system.

Description

Technical Field

[0001] This application relates to the field of land remote sensing monitoring technology, and in particular to a near real-time land degradation monitoring method and system. Background Technology

[0002] With the rapid growth of the world's population and the increasing severity of climate change, the conflict between humans and land is gradually intensifying, and land degradation has become one of the most serious global environmental problems. Sustainable use of land resources and monitoring of land degradation have become important issues. "Achieving a world with zero net land degradation" is one of the important goals of sustainable development, and carrying out large-scale near-real-time land degradation monitoring is of great significance to promoting sustainable development goals.

[0003] Generally, land degradation encompasses various forms of degradation processes. Currently, the definition of land degradation in the United Nations Convention to Combat Desertification (UNCCD) is widely accepted by countries as the reduction or loss of both the biological and economic productivity and complexity of land. UNCCD has published the "SDG 15.3.1 Good Practice Guidelines," which provides explanations of the definitions, calculation methods, and recommended data sources for the three core indicators of land degradation monitoring (land cover, land productivity, and soil carbon). This establishes a highly operational framework for land degradation monitoring and assessment, promoting the further development of large-scale land degradation monitoring research.

[0004] Existing technologies include many relatively mature large-scale land degradation monitoring systems, which have yielded corresponding monitoring results at the national level, and related results at the global level have also emerged in recent years. However, because remote sensing data requires further investigation, processing, and annotation before it can be utilized, most existing large-scale land degradation monitoring systems are developed based on historical remote sensing data, resulting in insufficient utilization of the real-time nature of remote sensing data. This leads to a certain lag in land degradation monitoring results, making it difficult to provide a timely and accurate data foundation for land health early warning, land use policy formulation, and sustainable development research and decision-making. Summary of the Invention

[0005] To address the problems existing in the prior art, this application provides a near real-time land degradation monitoring method and system, which can at least partially solve the problems existing in the prior art.

[0006] Firstly, this application provides a near-real-time land degradation monitoring method, including:

[0007] Historical and near-real-time multi-source remote sensing data of the target area are acquired and spatiotemporal sequence reconstruction is performed to obtain the spectral feature time series of each pixel in the target area. The spectral feature time series includes the near-real-time spectral feature time series and the historical spectral feature time series of different periods.

[0008] The terrain and climate features of the target area are obtained and fused with the time series of the spectral features to obtain a multidimensional feature vector for each pixel.

[0009] Obtain the historical land cover classification of each pixel in the first historical reference period, and calculate the near real-time land cover classification of each pixel based on the historical land cover classification and the multi-dimensional feature vector.

[0010] Based on the multidimensional feature vector of each pixel, and / or historical land cover classification and near real-time land cover classification, calculate the land degradation assessment index of each pixel, and assess the land degradation status of the pixel.

[0011] The multi-source remote sensing data includes high spatial and low temporal resolution remote sensing data and low spatial and high temporal resolution remote sensing data, and the spectral feature time series is a high spatial and high temporal resolution spectral feature time series.

[0012] The multidimensional feature vector includes: a historical multidimensional feature vector and a near-real-time multidimensional feature vector; the calculation of the near-real-time land cover classification for each pixel based on the historical land cover classification and the multidimensional feature vector includes:

[0013] Randomly select pixels as sample points, and calculate the change vector of the sample points based on the historical multidimensional feature vector and the near real-time multidimensional feature vector of the sample points in the first historical reference period.

[0014] The historical land cover classification of sample points whose angle and length of the change vector are both less than a preset first threshold are migrated to the near real-time land cover classification of the corresponding sample points to obtain a migration sample set.

[0015] The pre-established land cover classification model was trained using the aforementioned migration sample set;

[0016] The near-real-time multidimensional feature vectors of each pixel in the target region are input into the trained land cover classification model to obtain the near-real-time land cover classification of each pixel.

[0017] The multidimensional feature vector includes historical multidimensional feature vectors and near-real-time multidimensional feature vectors; the land degradation assessment indicators include historical values ​​of vegetation index and vegetation productivity index, and near-real-time values ​​of vegetation index and vegetation productivity index; the calculation of land degradation assessment indicators for each pixel based on the multidimensional feature vectors of each pixel, and / or historical land cover classification and near-real-time land cover classification, and the assessment of the land degradation status of the pixels, includes:

[0018] The historical index values ​​of vegetation index and vegetation productivity index are calculated based on the historical multidimensional feature vector.

[0019] The near-real-time index values ​​of the vegetation index and vegetation productivity index are calculated based on the near-real-time multidimensional feature vector.

[0020] Calculate the mean and variance of each historical value of the vegetation index and vegetation productivity index;

[0021] Determine whether the mean of each near-real-time indicator value of the vegetation index and vegetation productivity index differs from the mean of the corresponding historical indicator values ​​by more than the variance. If so, the land is considered to be degraded.

[0022] The land degradation assessment index includes a land cover change sub-index; the calculation of the land degradation assessment index for each pixel based on the multidimensional feature vector of each pixel and / or historical land cover classification and near-real-time land cover classification, and the assessment of the land degradation status of the pixel, includes:

[0023] By comparing the historical land cover classification with the near real-time land cover classification, the land cover change sub-index of the pixel is obtained according to the preset land cover transformation matrix. The land cover change sub-index represents the degradation of land cover vegetation.

[0024] This also includes:

[0025] Select any one of the vegetation index and vegetation productivity index as the first vegetation index, and perform regression analysis on the historical index value and near real-time index value of the first vegetation index of each pixel to obtain the slope and significance value of the linear regression of each pixel.

[0026] The land degradation trend of each pixel is determined based on the slope, significance value, and preset second threshold.

[0027] Select any one of the vegetation index and vegetation productivity index as the second vegetation index, calculate the mean and standard deviation of the historical index value of the second vegetation index of each pixel in the second historical reference period, and the mean of the historical index value and the near real-time index value of the second vegetation index of each pixel in the monitoring target period.

[0028] The significance of the second vegetation index is calculated based on the mean and standard deviation of the second historical baseline period and the mean of the monitoring target period;

[0029] The land degradation status of each pixel is determined based on the significance of the second vegetation index and a preset third threshold.

[0030] Select any one of the vegetation index and vegetation productivity index as the third vegetation index, divide the monitoring area into multiple target zones, and calculate the land degradation performance value of each pixel based on the maximum value of the near real-time index value of the third vegetation index of each pixel in each target zone and the near real-time index value of the third vegetation index of each pixel.

[0031] The land degradation performance of each pixel is determined based on the land degradation performance value and the preset fourth threshold.

[0032] Determine whether the land degradation trend, land degradation state, and land degradation performance of each pixel are degraded. If both the land degradation state and the land degradation performance are degraded and / or the land degradation trend is degraded, then the land productivity change sub-index of the corresponding pixel is degraded; otherwise, the land productivity change sub-index of the corresponding pixel is not degraded.

[0033] The multidimensional feature vector includes historical multidimensional feature vectors and near-real-time multidimensional feature vectors; the land degradation assessment index includes a land productivity change sub-index; the calculation of the land degradation assessment index for each pixel based on the multidimensional feature vector of each pixel and / or historical land cover classification and near-real-time land cover classification, and the assessment of the land degradation status of the pixel, includes:

[0034] Calculate the historical index values ​​of the vegetation index and vegetation productivity index based on the historical multidimensional feature vector;

[0035] The near-real-time index values ​​of the vegetation index and vegetation productivity index are calculated based on the near-real-time multidimensional feature vector.

[0036] Select any one of the vegetation index and vegetation productivity index as the first vegetation index, and perform regression analysis on the historical index value and near real-time index value of the first vegetation index of each pixel to obtain the slope and significance value of the linear regression of each pixel.

[0037] The land degradation trend of each pixel is determined based on the slope, significance value, and preset fifth threshold.

[0038] Select any one of the vegetation index and vegetation productivity index as the second vegetation index, calculate the mean and standard deviation of the historical index value of the second vegetation index of each pixel in the second historical reference period, and the mean of the historical index value and the near real-time index value of the second vegetation index of each pixel in the monitoring target period.

[0039] The significance of the second vegetation index is calculated based on the mean and standard deviation of the second historical baseline period and the mean of the monitoring target period;

[0040] The land degradation status of each pixel is determined based on the significance of the second vegetation index and the preset sixth threshold.

[0041] Select any one of the vegetation index and vegetation productivity index as the third vegetation index, divide the monitoring area into multiple target zones, and calculate the land degradation performance value of each pixel based on the maximum value of the near real-time index value of the third vegetation index of each pixel in each target zone and the near real-time index value of the third vegetation index of each pixel.

[0042] The land degradation performance of each pixel is determined based on the land degradation performance value and the preset seventh threshold.

[0043] Determine whether the land degradation trend, land degradation state, and land degradation performance of each pixel are degraded. If both the land degradation state and the land degradation performance are degraded and / or the land degradation trend is degraded, then the land productivity change sub-index of the corresponding pixel is degraded; otherwise, the land productivity change sub-index of the corresponding pixel is not degraded.

[0044] This also includes:

[0045] Obtain the near-real-time soil carbon estimate of the pixel and the soil carbon estimate of the historical baseline year. Calculate the soil carbon change sub-index of the pixel based on the near-real-time soil carbon estimate of the pixel, the soil carbon estimate of the historical baseline year, and the difference between the current year and the historical baseline year.

[0046] The soil carbon degradation status of the pixel is determined based on the soil carbon change sub-indicator and the preset threshold range.

[0047] This also includes:

[0048] Obtain the historical and near-real-time vegetation health indices of pixels;

[0049] Calculate the mean and variance of each indicator value in the historical vegetation health index;

[0050] Determine whether the difference between the mean value of each indicator in the near-real-time vegetation health index and the mean value of the corresponding historical indicator exceeds the variance. If so, the land is considered to be degraded.

[0051] Secondly, this application provides a near-real-time land degradation monitoring system, comprising:

[0052] The time series reconstruction unit is used to acquire historical and near-real-time multi-source remote sensing data of the target area and perform spatiotemporal sequence reconstruction to obtain the spectral feature time series of each pixel in the target area. The spectral feature time series includes the near-real-time spectral feature time series and the historical spectral feature time series of different periods.

[0053] A multi-dimensional feature fusion unit is used to acquire the terrain and climate features of the target area and fuse them with the time series of the spectral features to obtain the multi-dimensional feature vector of each pixel.

[0054] The land cover classification calculation unit is used to obtain the historical land cover classification of each pixel in the first historical reference period, and to calculate the near real-time land cover classification of each pixel based on the historical land cover classification and the multi-dimensional feature vector.

[0055] The land degradation assessment unit is used to calculate the land degradation assessment index of each pixel based on the multidimensional feature vector of each pixel and / or historical land cover classification and near real-time land cover classification, and to assess the land degradation status of the pixel.

[0056] Thirdly, this application provides a computer electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method described in any of the above embodiments.

[0057] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the methods described in any of the above embodiments.

[0058] Fifthly, this application provides a computer program product, which includes a computer program that, when executed by a processor, implements the method described in any of the above embodiments.

[0059] The near-real-time land degradation monitoring method and system provided in this application acquires historical and near-real-time multi-source remote sensing data of the target area and performs spatiotemporal sequence reconstruction to obtain the spectral feature time series of each pixel in the target area; acquires the topographic and climatic features of the target area and fuses them with the spectral feature time series to obtain the multidimensional feature vector of each pixel; acquires the historical land cover classification of each pixel in the first historical reference period, and calculates the near-real-time land cover classification of each pixel based on the historical land cover classification and the multidimensional feature vector; calculates the land degradation assessment index of each pixel based on the multidimensional feature vector and / or the historical land cover classification and the near-real-time land cover classification, and assesses the land degradation status of the pixels. This enables rapid prediction and utilization of land cover classification from near-real-time data, ensures the real-time nature of monitoring results, and fully utilizes various near-real-time data to calculate degradation indicators of the monitored area in multiple aspects such as land vegetation cover, land productivity, and soil carbon, realizing the monitoring of multiple aspects of land degradation, forming a complete monitoring system, and enhancing the accuracy and completeness of monitoring results. Attached Figure Description

[0060] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0061] Figure 1 This is a flowchart of a near-real-time land degradation monitoring method provided in an embodiment of this application;

[0062] Figure 2 This is a flowchart of a near-real-time land degradation monitoring method provided in an embodiment of this application;

[0063] Figure 3 This is a flowchart of a near-real-time land degradation monitoring method provided in an embodiment of this application;

[0064] Figure 4 This is a flowchart of a near-real-time land degradation monitoring method provided in an embodiment of this application;

[0065] Figure 5 This is a flowchart of a near-real-time land degradation monitoring method provided in an embodiment of this application;

[0066] Figure 6 This is a flowchart of a near-real-time land degradation monitoring method provided in an embodiment of this application;

[0067] Figure 7 This is a flowchart of a near-real-time land degradation monitoring method provided in an embodiment of this application;

[0068] Figure 8 This is a flowchart of a near-real-time land degradation monitoring method provided in an embodiment of this application;

[0069] Figure 9 This is a schematic diagram of the structure of a near real-time land degradation monitoring system provided in one embodiment of this application;

[0070] Figure 10 This is a schematic diagram of the structure of a near real-time land degradation monitoring system provided in one embodiment of this application;

[0071] Figure 11 This is a schematic diagram of the structure of a near real-time land degradation monitoring system provided in one embodiment of this application;

[0072] Figure 12 This is a schematic diagram of the structure of a near real-time land degradation monitoring system provided in one embodiment of this application;

[0073] Figure 13 This is a schematic diagram of the structure of a near real-time land degradation monitoring system provided in one embodiment of this application;

[0074] Figure 14 This is a schematic diagram of the structure of a near real-time land degradation monitoring system provided in one embodiment of this application;

[0075] Figure 15 This is a schematic diagram of the structure of a near real-time land degradation monitoring system provided in one embodiment of this application;

[0076] Figure 16 This is a schematic diagram of the structure of a near real-time land degradation monitoring system provided in one embodiment of this application;

[0077] Figure 17 This is a schematic diagram of the physical structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0078] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings. Here, the illustrative embodiments and descriptions of the present invention are used to explain the present invention, but are not intended to limit the present invention. It should be noted that, unless otherwise specified, the embodiments and features in the embodiments of this application can be arbitrarily combined with each other.

[0079] The following describes the specific implementation process of the near real-time land degradation monitoring method with code synchronization provided in this embodiment of the invention, using a server as the execution subject as an example.

[0080] Figure 1 This is a flowchart of a near-real-time land degradation monitoring method provided in an embodiment of this application, such as... Figure 1As shown, the near-real-time land degradation monitoring method provided in this application includes:

[0081] S101: Acquire historical and near-real-time multi-source remote sensing data of the target area and perform spatiotemporal sequence reconstruction to obtain the spectral feature time series of each pixel in the target area. The spectral feature time series includes the near-real-time spectral feature time series and the historical spectral feature time series of different periods.

[0082] S102: Obtain the terrain and climate features of the target area and fuse them with the time series of spectral features to obtain the multidimensional feature vector of each pixel;

[0083] S103: Obtain the historical land cover classification of each pixel in the first historical reference period, and calculate the near real-time land cover classification of each pixel based on the historical land cover classification and multi-dimensional feature vector.

[0084] S104: Calculate the land degradation assessment index for each pixel based on the multidimensional feature vector of each pixel and / or historical land cover classification and near real-time land cover classification, and assess the land degradation status of the pixel.

[0085] The near-real-time land degradation monitoring method and system provided in this application acquires historical and near-real-time multi-source remote sensing data of the target area and performs spatiotemporal sequence reconstruction to obtain the spectral feature time series of each pixel in the target area; acquires the topographic and climatic features of the target area and fuses them with the spectral feature time series to obtain the multidimensional feature vector of each pixel; acquires the historical land cover classification of each pixel in the first historical reference period, and calculates the near-real-time land cover classification of each pixel based on the historical land cover classification and the multidimensional feature vector; calculates the land degradation assessment index of each pixel based on the multidimensional feature vector and / or the historical land cover classification and the near-real-time land cover classification, and assesses the land degradation status of the pixels. This enables rapid prediction and utilization of land cover classification from near-real-time data, ensures the real-time nature of monitoring results, and fully utilizes various near-real-time data to calculate degradation indicators of the monitored area in multiple aspects such as land vegetation cover, land productivity, and soil carbon, realizing the monitoring of multiple aspects of land degradation, forming a complete monitoring system, and enhancing the accuracy and completeness of monitoring results.

[0086] The following is a detailed explanation of each step.

[0087] S101: Acquire historical and near-real-time multi-source remote sensing data of the target area and perform spatiotemporal sequence reconstruction to obtain the spectral feature time series of each pixel in the target area. The spectral feature time series includes the near-real-time spectral feature time series and the historical spectral feature time series of different periods.

[0088] Specifically, due to factors such as weather and equipment limitations, for example, some satellites may be obscured by clouds during cloudy weather, resulting in data loss. The server can use multi-source remote sensing data to fill in the missing data, obtaining more complete data, improving resolution, and completing spatiotemporal sequence reconstruction to obtain the spectral feature time series of each pixel in the target area. Spatiotemporal sequence reconstruction includes the reconstruction of historical multi-source remote sensing data and the reconstruction of near-real-time multi-source remote sensing data to obtain near-real-time spectral feature time series and historical spectral feature time series from different periods.

[0089] Based on the above embodiments, the multi-source remote sensing data further includes high spatial and low temporal resolution remote sensing data and low spatial and high temporal resolution remote sensing data, and the spectral feature time series is a high spatial and high temporal resolution spectral feature time series.

[0090] Specifically, in existing remote sensing data, those with high spatial resolution often have low temporal resolution, and those with high temporal resolution often have low spatial resolution. Therefore, the server fuses multi-source remote sensing data through spatiotemporal sequence reconstruction to obtain a spectral feature time series with high spatial and temporal resolution. Appropriate spatiotemporal sequence reconstruction methods can be selected based on actual conditions, and this application does not impose any restrictions on this.

[0091] For example, such as Figure 2 As shown, spatiotemporal sequence reconstruction includes:

[0092] S201: Preprocess the multi-source remote sensing data to obtain the first projection data and the second projection data.

[0093] Specifically, the preprocessing includes: selecting high-quality, low-temporal-resolution remote sensing data and reprojecting it to obtain first projected data; selecting low-quality, high-temporal-resolution remote sensing data and reprojecting it using bicubic interpolation to obtain second projected data. The first projected data is high-spatial-low-temporal-resolution projected data, and the second projected data is low-spatial-high-temporal-resolution projected data. Specific quality selection criteria can be set according to specific circumstances, such as the proportion of influence from cloud cover, etc., which this application does not limit. Through preprocessing, low-quality remote sensing data can be removed, and multi-source remote sensing data can be mapped to the same coordinate system through reprojection.

[0094] S202: Synthesize the first projection data and the second projection data respectively to obtain the first spectral feature time series and the second spectral feature time series;

[0095] Specifically, both the first projection data and the second projection data contain projections of remote sensing image data from multiple sources. Therefore, they are synthesized according to preset time intervals to obtain the first spectral feature time series and the second spectral feature time series.

[0096] S203: Construct a conversion model based on the first spectral feature time series and the second spectral feature time series;

[0097] Specifically, the server overlays the first spectral feature time series and the second spectral feature time series to obtain the overlapping region between the first spectral feature time series and the second spectral feature time series. Sample points are randomly selected in the overlapping region, and the values ​​of the first spectral feature time series and the second spectral feature time series corresponding to the sample points are used as the two attribute values ​​of the sample points. A linear regression model is constructed based on the attribute values ​​of all sample points as the conversion model between the first spectral feature time series and the second spectral feature time series.

[0098] S204: Supplement missing values ​​in the first spectral feature time series based on the second spectral feature time series and the transformation model;

[0099] Specifically, the server acquires the time points in the first spectral feature time series where data is missing at preset time intervals, and calculates the missing values ​​based on the values ​​of the second spectral feature time series corresponding to each missing time point and the transformation model, thus supplementing the first spectral feature time series to obtain a spectral feature time series with high temporal and spatial resolution. Furthermore, the obtained high temporal and spatial resolution spectral feature time series can be further smoothed and denoised.

[0100] In addition, spatiotemporal sequence reconstruction can also be performed using the Spatial and Temporal Adaptive Reflectance Fusion Model (STARFM) and the Enhanced Spatialand and Temporal Adaptive Reflectance Fusion Model (ESTARFM), and this application does not impose any restrictions on this.

[0101] Spatiotemporal sequence reconstruction can fully utilize remote sensing data from multiple sources, fill in missing data, improve data accuracy, and complete the fusion of high spatial and low temporal resolution remote sensing data and low spatial and high temporal resolution remote sensing data to obtain high spatial and high temporal resolution remote sensing data, thereby further improving the accuracy of subsequent land degradation monitoring and analysis.

[0102] S102: Obtain the terrain and climate features of the target area and fuse them with the time series of spectral features to obtain the multidimensional feature vector of each pixel;

[0103] Specifically, by using various relevant data statistics websites, publicly available data on the topographic and climatic features of the target area at different times are obtained and fused with the spectral feature time series to obtain a multidimensional feature vector for each pixel.

[0104] S103: Obtain the historical land cover classification of each pixel in the first historical reference period, and calculate the near real-time land cover classification of each pixel based on the historical land cover classification and multi-dimensional feature vector.

[0105] Specifically, an appropriate year can be selected as the first base year based on the actual situation. For example, the first historical base year can be selected with reference to the recommendations of the United Nations Convention to Combat Desertification (UNCCD). This application does not impose any restrictions on this.

[0106] Based on the above embodiments, the multidimensional feature vector further includes: historical multidimensional feature vector and near real-time multidimensional feature vector; such as Figure 3 As shown, the near-real-time land cover classification of each pixel is calculated based on historical land cover classification and multi-dimensional feature vectors, including:

[0107] S301: Randomly select pixels as sample points, and calculate the change vector of the sample points based on the historical multidimensional feature vector of the first historical benchmark period and the near real-time multidimensional feature vector of the sample points.

[0108] Specifically, the change vector can be obtained by subtracting the historical multidimensional feature vector of the first historical reference period from the near-real-time multidimensional feature vector.

[0109] S302: Transfer the historical land cover classification of sample points whose angle and length of the change vector are both less than the preset first threshold to the near real-time land cover classification of the corresponding sample points to obtain the transferred sample set;

[0110] Specifically, the angle and magnitude of the change vector reflect the degree of change; the smaller the angle and magnitude, the smaller the degree of change. A first threshold for the angle and length can be set according to the actual situation. This allows for the migration of historical land cover classifications of sample points whose angle and length of the change vector are both less than the preset first threshold to the corresponding near-real-time land cover classifications. The smaller the first threshold, the higher the accuracy of the sample migration.

[0111] S303: Train a pre-established land cover classification model using a migration sample set;

[0112] Specifically, the land cover classification model can select a suitable classification algorithm model based on the actual situation, such as K-means clustering or random forest, and use the migration sample set obtained in S302 for training. The migration sample set can be divided into a training set and a prediction set as needed.

[0113] S304: Input the near real-time multidimensional feature vectors of each pixel in the target region into the trained land cover classification model to obtain the near real-time land cover classification of each pixel.

[0114] Specifically, the server inputs the near real-time multidimensional feature vectors of each pixel in the target area into the trained land cover classification model, and uses the trained land cover classification model to make predictions to obtain the near real-time land cover classification of each pixel.

[0115] By randomly selecting sample points, the change vector of the sample points is calculated; sample points whose angle and length of the change vector are both less than a preset first threshold are subjected to land cover classification migration to obtain a migration sample set; a pre-established land cover classification model is trained using the migration sample set; the near real-time multidimensional feature vector of each pixel in the target area is input into the trained land cover classification model to obtain the near real-time land cover classification of each pixel. This achieves rapid prediction of land cover classification based on near real-time data, saving a significant amount of time in obtaining land cover labels through surveys and statistics, thereby realizing the utilization of near real-time data and enhancing the real-time nature of monitoring results.

[0116] S104: Calculate the land degradation assessment index for each pixel based on the multidimensional feature vector of each pixel and / or historical land cover classification and near real-time land cover classification, and assess the land degradation status of the pixel.

[0117] Specifically, the multidimensional feature vector of each pixel contains the reflectance of various bands in the spectral feature time series, such as the reflectance of red light, blue light, and near-infrared bands, as well as various topographic and climatic features. Based on these parameters, multiple land assessment indicators for the target area can be calculated, and the land can be judged as to whether it has degraded in the corresponding aspects based on these indicators.

[0118] Based on the above embodiments, the multidimensional feature vector further includes: a historical multidimensional feature vector and a near-real-time multidimensional feature vector; the land degradation assessment indicators include historical values ​​of vegetation index and vegetation productivity index, and near-real-time values ​​of vegetation index and vegetation productivity index; such as Figure 4 As shown, based on the multidimensional feature vector of each pixel and / or historical land cover classification and near-real-time land cover classification, the land degradation assessment index of each pixel is calculated, and the land degradation status of the pixels is assessed, including:

[0119] S401: Calculate the historical values ​​of vegetation index and vegetation productivity index based on historical multidimensional feature vectors;

[0120] Specifically, vegetation indices may include the Normalized Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), Normalized Burn Ratio (NBR), and Bare Soil Index (BSI), while vegetation productivity indices may include Gross Primary Productivity (GPP) and Net Primary Productivity (NPP). These indicators can be calculated using the following formulas:

[0121]

[0122]

[0123]

[0124]

[0125] NPP = CPP - Rh (5)

[0126] Wherein, NIR is the reflectance of the pixel in the near-infrared band, R is the reflectance of the pixel in the red band, SWIR is the reflectance of the pixel in the short-infrared band, B is the reflectance of the pixel in the blue band, C1, C2, and L are constants that vary depending on the remote sensing data source used, and Rh is the respiration consumption of heterotrophic organisms. Furthermore, GPP can be estimated using appropriate methods depending on the specific circumstances such as the detection area. For example, it can be estimated using the VPM model with EVI data, or using the MODIS / GPP algorithm, etc. This application does not impose any limitations on this. Using the above formulas and methods, the historical values ​​of vegetation indices and vegetation productivity indices can be calculated based on historical multidimensional feature vectors. Since the historical multidimensional feature vectors contain multiple periods, each historical indicator value includes multiple values ​​from different periods.

[0127] S402: Calculate the near-real-time values ​​of vegetation index and vegetation productivity index based on near-real-time multidimensional feature vectors;

[0128] Specifically, the near-real-time values ​​of vegetation index and vegetation productivity index can also be calculated based on the near-real-time multidimensional feature vector and formulas (1)-(5).

[0129] S403: Calculate the mean and variance of the historical values ​​of the vegetation index and the vegetation productivity index.

[0130] Specifically, the mean and variance of multiple values ​​from different periods for each historical indicator of vegetation index and vegetation productivity index are calculated as a benchmark to determine whether each near-real-time indicator value is within the normal range.

[0131] S404: Determine whether the mean of each near-real-time indicator value of the vegetation index and vegetation productivity index differs from the mean of the corresponding historical indicator values ​​by more than the variance. If so, the land is considered to be degraded.

[0132] Specifically, it is determined whether the difference between the mean and the corresponding historical value of each near-real-time indicator of the vegetation index and vegetation productivity index exceeds the variance, that is, whether each near-real-time indicator value is within the range of [Mean-Std, Mean+Std], where Mean is the mean of the corresponding historical indicator value and Std is the variance of the corresponding historical indicator value. If the difference between a near-real-time indicator value and the mean of the corresponding historical indicator value exceeds the variance, that is, the near-real-time indicator is not within the normal range, it indicates that the land may be degraded in terms of vegetation health.

[0133] By calculating the historical and near-real-time values ​​of vegetation index and vegetation productivity index based on historical and near-real-time multidimensional feature vectors, and calculating the near-real-time values ​​of vegetation index and vegetation productivity index based on near-real-time multidimensional feature vectors, it is possible to determine whether the difference between the mean of the near-real-time values ​​of vegetation index and vegetation productivity index and the corresponding historical values ​​exceeds the variance. This enables the detection of land vegetation health through multiple indicators and the determination of whether there is land degradation in terms of vegetation health.

[0134] Based on the above, the land degradation assessment index further includes a land cover change sub-index; the land degradation assessment index for each pixel is calculated based on the multidimensional feature vector of each pixel, and / or historical land cover classification and near-real-time land cover classification, and the land degradation status of the pixels is assessed, including:

[0135] By comparing historical land cover classification with near-real-time land cover classification, land cover change sub-indices for pixels are obtained based on a preset land cover transformation matrix. These sub-indices represent the degradation of land cover vegetation.

[0136] Specifically, based on the predicted near-real-time land cover classification of each pixel, a comparison can be made with the corresponding historical land cover classification using Table 1 to obtain sub-indicators of land cover change for each pixel, thereby assessing the land degradation status of each pixel. For example, if a pixel's near-real-time land cover classification is grassland and its historical land cover classification is forest, then Table 1 shows that the sub-indicator of land cover change for this pixel is degradation, therefore it is considered that the land area corresponding to this pixel may be degraded.

[0137] Table 1 Land degradation assessment based on land cover transformation matrix

[0138]

[0139] By comparing historical land cover classification with near-real-time land cover classification, land cover change sub-indices are obtained for each pixel based on a preset land cover transformation matrix. These sub-indices represent the degradation of land cover vegetation, enabling the monitoring of land vegetation cover and thus assessing whether land vegetation cover is degraded.

[0140] exist Figure 4 Based on the embodiments, further, such as Figure 5 As shown, the near-real-time land degradation monitoring method provided in this application also includes:

[0141] S501: Select any one of the vegetation index and vegetation productivity index as the first vegetation index, and perform regression analysis on the historical and near real-time index values ​​of the first vegetation index for each pixel to obtain the slope and significance value of the linear regression for each pixel.

[0142] Specifically, the land degradation status of each pixel is assessed from three aspects: "trend," "status," and "performance." Among these, the "trend" aspect can be assessed from... Figure 4 In this embodiment, one of the vegetation indices and vegetation productivity indices calculated is arbitrarily selected as the first vegetation index. Regression analysis is performed based on multiple historical and near-real-time index values ​​of this index for each pixel to obtain the slope and significance value of the linear regression for each pixel, so as to obtain the trend of vegetation growth from the historical period to the current monitoring period for each pixel.

[0143] S502: Determine the land degradation trend of each pixel based on the slope, significance value, and preset second threshold;

[0144] Specifically, the slope and significance value can be compared with a preset second threshold to determine whether the land corresponding to each pixel has degraded. The second threshold can be set according to the selected regression analysis method, the historical slope and significance value of the monitoring area, etc., and this application does not impose any restrictions on it.

[0145] For example, ridge regression analysis can be used to monitor the trend of NDVI changes in pixels. If p < 0.01 and slope > 0 in the regression analysis results of a pixel, the trend of the pixel is considered to be recovery. If p < 0.01 and slope < 0 in the recovery analysis results of a pixel, the trend of the pixel is considered to be degradation. Otherwise, the trend of the pixel is considered to be stable. Here, p is the slope and slope is the significance value.

[0146] S503: Select any one of the vegetation index and vegetation productivity index as the second vegetation index, calculate the mean and standard deviation of the historical index value of the second vegetation index of each pixel in the second historical reference period, and the mean of the historical index value and the near real-time index value of the second vegetation index of each pixel in the monitoring target period.

[0147] Specifically, regarding the "state," it can be seen from... Figure 4 In this embodiment, one of the calculated vegetation indices and vegetation productivity indices is arbitrarily selected as the second vegetation index. The mean and standard deviation of the second vegetation index for each pixel in the second historical reference period are calculated, as well as the mean of the historical and near-real-time index values ​​of the second vegetation index for each pixel in the monitoring target period, to compare the vegetation growth status of the same pixel in the historical reference period and the current monitoring target period. The second historical period and the monitoring target period can be selected according to certain standards based on actual conditions. For example, the second historical reference period can be selected with reference to the recommendations of the United Nations Convention to Combat Desertification (UNCCD), and the period from the second historical reference period to the current year can be set as the monitoring target period. This application does not impose any restrictions on this.

[0148] S504: Calculate the significance of the second vegetation index based on the mean and standard deviation of the second historical baseline period and the mean of the monitoring target period;

[0149] Specifically, significance can be calculated using the following formula:

[0150]

[0151] Where Z represents significance. denoted as the mean of the second vegetation index during the target monitoring period, μ is the mean of the second vegetation index during the second historical baseline period, and σ is the standard deviation of the second vegetation index during the second historical baseline period.

[0152] S505: Determine the land degradation status of each pixel based on the significance of the second vegetation index and the preset third threshold.

[0153] Specifically, the settings can be made based on the actual situation such as the monitored area and its historical status. For example, using NDVI as the second vegetation index, if Z < -1.96, the state of the pixel is determined to be degraded; if Z > 1.96, the pixel is determined to be restored; otherwise, the state of the pixel is determined to be stable.

[0154] S506: Select any one of the vegetation index and vegetation productivity index as the third vegetation index, divide the monitoring area into multiple target zones, and calculate the land degradation performance value of each pixel based on the maximum value of the near real-time index value of the third vegetation index of each pixel in each target zone and the near real-time index value of the third vegetation index of each pixel.

[0155] Specifically, in terms of "performance", it can be seen from... Figure 4 In this embodiment, one of the calculated vegetation index and vegetation productivity index is arbitrarily selected as the third vegetation index. Based on global zoning data, the monitoring area is divided into multiple target zones. The near-real-time index value of the third vegetation index of a pixel is divided by the maximum value of the near-real-time index values ​​of the third vegetation index of all pixels in the target zone where the pixel is located, and the result is taken as the performance value of the pixel.

[0156] For example, if NPP is selected as the third vegetation indicator, the specific value of land degradation performance can be calculated according to the following formula:

[0157]

[0158] Where Performance is the performance value, and NPP is the performance value. observed NPP is the NPP value of a pixel. max This is the maximum value among the near-real-time values ​​of the third vegetation index of all pixels in the target partition where the pixel is located. To eliminate interference, the third vegetation index value located at the x% quantile of the ecological partition can be selected as the maximum value. x can be set according to the specific circumstances such as the selected index and the monitored area, for example, 90%. This application does not impose any restrictions on this.

[0159] S507: Determine the land degradation performance of each pixel based on the land degradation performance value and the preset fourth threshold;

[0160] Specifically, the land degradation performance value is compared with a preset fourth threshold to determine whether each pixel exhibits degradation. The fourth threshold and the specific comparison method can be set according to the actual situation, such as the selected parameters. For example, when NPP is selected as the third vegetation index value, the fourth threshold is set to 0.5. When the performance of a pixel is less than 0.5, it is considered that the area corresponding to that pixel exhibits land degradation.

[0161] S508: Determine whether the land degradation trend, land degradation state, and land degradation performance of each pixel are degraded. If the land degradation state and the land degradation performance are both degraded and / or the land degradation trend is degraded, then the land productivity change sub-index of the corresponding pixel is degraded; otherwise, the land productivity change sub-index of the corresponding pixel is not degraded.

[0162] Specifically, based on the "trend," "status," and "performance" of each pixel, we can comprehensively determine whether the land productivity change sub-indicator of the area corresponding to that pixel has degraded. The following table can be used as a reference for this determination:

[0163] Table 2. Land Productivity Degradation Monitoring Finding Table

[0164] trend state Performance Combining the three Y Y Y Y Y Y N Y Y N Y Y Y N N Y N Y Y Y N Y N N N N Y N N N N N

[0165] Where Y represents degradation and N represents non-degradation, that is, when the land degradation state and the land degradation manifestation are both degradation and / or the land degradation trend is degradation, the land productivity of the area corresponding to the pixel is considered to be degraded.

[0166] By performing regression analysis on the vegetation index and vegetation productivity index of each pixel, calculating significance and performance values, and comparing them with the corresponding thresholds, it was determined whether each pixel had degraded in terms of "trend", "state" and "performance". A comprehensive analysis of the three indicators further determined whether the pixel had degraded in terms of land productivity, forming a relatively complete land degradation monitoring system. This system made full use of remote sensing monitoring data and intermediate data, resulting in more professional land degradation monitoring results.

[0167] exist Figure 1 Based on the embodiments, further, such as Figure 6 As shown, the multidimensional feature vector includes: historical multidimensional feature vector and near-real-time multidimensional feature vector; the land degradation assessment index includes the land productivity change sub-index; based on the multidimensional feature vector of each pixel, and / or historical land cover classification and near-real-time land cover classification, the land degradation assessment index of each pixel is calculated, and the land degradation status of the pixel is assessed, including:

[0168] S601: Calculate the historical values ​​of vegetation index and vegetation productivity index based on historical multidimensional feature vectors;

[0169] Specifically, vegetation indices may include the Normalized Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), Normalized Burn Ratio (NBR), and Bare Soil Index (BSI), etc. Vegetation productivity indices may include Gross Primary Productivity (GPP) and Net Primary Productivity (NPP), etc. The historical index values ​​of the above indicators are calculated based on the historical multidimensional feature vectors, and can be specifically calculated according to formulas (1)-(5).

[0170] S602: Calculate the near-real-time values ​​of vegetation index and vegetation productivity index based on near-real-time multidimensional feature vectors;

[0171] Specifically, the near-real-time index values ​​of vegetation index and vegetation productivity index of each pixel are calculated using near-real-time multidimensional feature vectors according to formulas (1)-(5).

[0172] S603: Select any one of the vegetation index and vegetation productivity index as the first vegetation index, and perform regression analysis on the historical and near real-time index values ​​of the first vegetation index for each pixel to obtain the slope and significance value of the linear regression for each pixel.

[0173] Specifically, regarding the "trend", one of the vegetation indices and vegetation productivity indices calculated from S601 and S602 can be arbitrarily selected as the first vegetation index. Regression analysis is performed based on multiple historical and near-real-time index values ​​of this index for each pixel to obtain the slope and significance value of the linear regression for each pixel, so as to obtain the trend of vegetation growth from the historical period to the current monitoring period for each pixel.

[0174] S604: Determine the land degradation trend of each pixel based on the slope, significance value, and preset fifth threshold;

[0175] Specifically, the slope and significance value can be compared with a preset fifth threshold to determine whether the land corresponding to each pixel has degraded. The fifth threshold can be set according to the selected regression analysis method, the historical slope and significance value of the monitoring area, etc., and this application does not impose any restrictions on it.

[0176] For example, ridge regression analysis can be used to monitor the trend of NDVI changes in pixels. If p < 0.01 and slope > 0 in the regression analysis results of a pixel, the trend of the pixel is considered to be recovery. If p < 0.01 and slope < 0 in the recovery analysis results of a pixel, the trend of the pixel is considered to be degradation. Otherwise, the trend of the pixel is considered to be stable. Here, p is the slope and slope is the significance value.

[0177] S605: Select any one of the vegetation index and vegetation productivity index as the second vegetation index, calculate the mean and standard deviation of the historical index value of the second vegetation index of each pixel in the second historical reference period, and the mean of the historical index value and the near real-time index value of the second vegetation index of each pixel in the monitoring target period.

[0178] Specifically, regarding the "status," one of the vegetation indices and vegetation productivity indices calculated from S601 and S602 can be arbitrarily selected as the second vegetation index. The mean and standard deviation of the second vegetation index for each pixel in the second historical reference period, as well as the mean of the historical and near-real-time index values ​​of the second vegetation index for each pixel in the target monitoring period, are calculated to compare the vegetation growth status of the same pixel in the historical reference period and the current target monitoring period. The second historical period and the target monitoring period can be selected according to certain standards based on actual conditions. For example, the second historical reference period can be selected with reference to the recommendations of the United Nations Convention to Combat Desertification (UNCCD), and the period from the second historical reference period to the current year can be set as the target monitoring period. This application does not impose any restrictions on this.

[0179] S606: Calculate the significance of the second vegetation index based on the mean and standard deviation of the second historical baseline period and the mean of the monitoring target period;

[0180] Specifically, significance can be calculated using formula (6).

[0181] S607: Determine the land degradation status of each pixel based on the significance of the second vegetation index and the preset sixth threshold.

[0182] Specifically, the sixth threshold can be set according to the actual situation such as the monitored area and the historical state of the area. For example, using NDVI as the second vegetation index, if Z < -1.96, the state of the pixel is determined to be degraded; if Z > 1.96, the pixel is determined to be restored; otherwise, the state of the pixel is determined to be stable. Z is the significance value of the second vegetation index of the pixel.

[0183] S608: Select any one of the vegetation index and vegetation productivity index as the third vegetation index, divide the monitoring area into multiple target zones, and calculate the land degradation performance value of each pixel based on the maximum value of the near real-time index value of the third vegetation index of each pixel in each target zone and the near real-time index value of the third vegetation index of each pixel.

[0184] Specifically, in terms of "performance", it can be seen from... Figure 4 In this embodiment, either the calculated vegetation index or the vegetation productivity index is arbitrarily selected as the third vegetation index. Based on global zoning data, the monitoring area is divided into multiple target zones. The near-real-time value of the third vegetation index of a pixel is divided by the maximum near-real-time value of the third vegetation index among all pixels in the target zone containing that pixel. The result is used as the performance value of that pixel. To eliminate interference, the maximum value can be the third vegetation index value located at the x% quantile of the ecological zone. x can be set according to the selected index, the monitored area, and other specific circumstances, for example, 90%. This application does not impose any restrictions on this.

[0185] S609: Determine the land degradation performance of each pixel based on the land degradation performance value and the preset seventh threshold;

[0186] Specifically, the land degradation performance value is compared with a preset seventh threshold to determine whether each pixel exhibits degradation. The seventh threshold and the specific comparison method can be set according to the selected parameters and other actual conditions. For example, when NPP is selected as the third vegetation index value, the seventh threshold is set to 0.5. When the performance of a pixel is less than 0.5, it is considered that the area corresponding to that pixel exhibits land degradation.

[0187] S610: Determine whether the land degradation trend, land degradation state, and land degradation performance of each pixel are degraded. If the land degradation state and the land degradation performance are both degraded and / or the land degradation trend is degraded, then the land productivity change sub-index of the corresponding pixel is degraded; otherwise, the land productivity change sub-index of the corresponding pixel is not degraded.

[0188] Specifically, based on the "trend", "state" and "performance" of each pixel, it is determined whether the land productivity change sub-indicator of the area corresponding to the pixel has degraded. Specifically, it can be determined according to Table 2. That is, when the land degradation state and the land degradation performance are both degraded and / or the land degradation trend is degraded, it is considered that the land productivity of the area corresponding to the pixel has degraded.

[0189] By calculating the vegetation index and vegetation productivity index of each pixel based on its historical and near-real-time multidimensional feature vectors, and performing regression analysis, significance and performance values, and comparing them with corresponding thresholds, we can determine whether each pixel has degraded in terms of "trend," "state," and "performance." By comprehensively analyzing the three indicators, we can further determine whether the pixel has degraded in terms of land productivity, thus forming a relatively complete land degradation monitoring system. This system makes full use of remote sensing monitoring data and intermediate data, resulting in more professional land degradation monitoring results.

[0190] In one embodiment, such as Figure 7 As shown, the near-real-time land degradation monitoring method provided in this application also includes:

[0191] S701: Obtain the near-real-time soil carbon estimate of the pixel and the soil carbon estimate of the historical baseline year. Calculate the soil carbon change sub-indicator of the pixel based on the near-real-time soil carbon estimate of the pixel, the soil carbon estimate of the historical baseline year, and the difference between the current year and the historical baseline year.

[0192] Specifically, the near-real-time soil carbon estimate for each pixel can be obtained from publicly available data sources. Alternatively, it can be fitted using the least squares method based on the reflectance of visible and near-infrared light in the historical multidimensional feature vector, historical soil carbon estimates, and soil carbon estimates for each corresponding historical year. The fitted curve is then used to calculate the near-real-time soil carbon estimate for each pixel based on the reflectance of visible and near-infrared light in the near-real-time multidimensional feature vector. The historical baseline year can be selected based on actual circumstances, such as choosing 15 years ago as the historical baseline year. This application does not impose any restrictions on comparisons. The soil carbon estimate for the historical baseline year can be obtained from publicly available data sources. The sub-indices of soil carbon change can be calculated using the following formula:

[0193]

[0194] Wherein, ΔSOC is a sub-indicator of soil carbon change, SOC0 is a near-real-time estimate of soil carbon, and SOC (0-T) The soil carbon estimate is for the historical baseline year, where c represents the ecological zone, s represents the soil type, p represents the pixel unit to be estimated, and SOC is... Ref For reference soil carbon values, F LU F is the stock change coefficient of a specific land use within a land use system or subsystem. MG F is a land management coefficient, representing different land management methods (such as different cultivation methods for arable land). I Let be the coefficient of change in the stock of organic matter input, A be the land area of ​​the ecological zone, and D be the difference between the current year and the historical baseline year.

[0195] S702: Determine the soil carbon degradation status of pixels based on soil carbon change sub-indicators and preset threshold ranges.

[0196] Specifically, the threshold range can be set and adjusted according to specific circumstances. In one embodiment, the correspondence between land cover change and soil carbon change proposed by the Intergovernmental Panel on Climate Change (IPCC) can be used to complete the assessment based on the calculated sub-indicators of soil carbon change, and finally divided into three levels: degradation, recovery and stability.

[0197] By calculating sub-indices of soil carbon change, the use of open-source data was incorporated into the land degradation assessment system, which yielded information on land degradation in terms of soil carbon. By combining this with the above-mentioned embodiments, the land degradation assessment system was further improved, the data sources and monitoring perspectives of land degradation were enriched, and a complete system for fully monitoring land degradation from multiple angles was realized.

[0198] In one embodiment, such as Figure 8 As shown, the near-real-time land degradation monitoring method provided in this application also includes:

[0199] S801: Obtain the historical and near-real-time vegetation health indices of pixels;

[0200] Specifically, vegetation health indices include the Agricultural Stress Index (ASI), Drought Intensity (DI), Mean Vegetation Health Index (MVHI), Vegetation Condition Index (VCI), and Vegetation Health Index (VHI). These data can be found through the open-source FAO-Agricultural Stress Index System (FAO-ASIS).

[0201] S802: Calculate the mean and variance of each indicator value in the historical vegetation health index;

[0202] Specifically, the historical vegetation health index includes the index values ​​of each indicator in different historical years, and calculates the mean and variance of each indicator in the historical vegetation health index in each historical year.

[0203] S803: Determine whether the difference between the mean value of each indicator in the near-real-time vegetation health index and the mean value of the corresponding historical indicator exceeds the variance. If so, it is considered that the land is degraded.

[0204] Specifically, the analysis determines whether the difference between the mean and the corresponding historical value of each near-real-time indicator of the vegetation health index exceeds the variance. That is, it determines whether each near-real-time indicator value is within the range of [Mean-Std, Mean+Std], where Mean is the mean of the corresponding historical indicator value and Std is the variance of the corresponding historical indicator value. If the difference between a near-real-time indicator value and the mean of the corresponding historical indicator value exceeds the variance, meaning that the near-real-time indicator is not within the normal range, it indicates that the land may be degraded in terms of vegetation health.

[0205] The near-real-time land degradation monitoring method provided in this application obtains the spectral feature time series of each pixel in the target area by acquiring historical and near-real-time multi-source remote sensing data of the target area and reconstructing the spatiotemporal sequence; it acquires the topographic and climatic features of the target area and fuses them with the spectral feature time series to obtain the multidimensional feature vector of each pixel; it acquires the historical land cover classification of each pixel in the first historical reference period, and calculates the near-real-time land cover classification of each pixel based on the historical land cover classification and the multidimensional feature vector; it calculates the land degradation assessment index of each pixel based on the multidimensional feature vector and / or the historical land cover classification and the near-real-time land cover classification, and assesses the land degradation status of the pixels. This method enables rapid prediction and utilization of land cover classification from near-real-time data, ensures the real-time nature of monitoring results, and fully utilizes various near-real-time data to calculate degradation indicators of the monitored area in multiple aspects such as land vegetation cover, land productivity, and soil carbon, achieving monitoring of multiple aspects of land degradation, forming a complete monitoring system, and enhancing the accuracy and completeness of monitoring results.

[0206] Based on the same inventive concept, this application also provides a near-real-time land degradation monitoring system, which can be used to implement the methods described in the above embodiments, as described in the following embodiments. Since the principle of the near-real-time land degradation monitoring system in solving the problem is similar to that of the near-real-time land degradation monitoring method, the implementation of the near-real-time land degradation monitoring system can refer to the implementation of the software performance benchmark determination method, and repeated details will not be elaborated further. As used below, the terms "unit" or "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the system described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0207] Figure 9 This is a schematic diagram of the structure of a near-real-time land degradation monitoring system provided in an embodiment of this application, as shown below. Figure 9 As shown, the near real-time land degradation monitoring system provided in this application includes:

[0208] The time series reconstruction unit 901 is used to acquire historical and near real-time multi-source remote sensing data of the target area and perform spatiotemporal sequence reconstruction to obtain the spectral feature time series of each pixel in the target area. The spectral feature time series includes the near real-time spectral feature time series and the historical spectral feature time series of different periods.

[0209] Specifically, due to factors such as weather and equipment limitations, for example, some satellites may be obscured by clouds during cloudy weather, resulting in data loss. The time series reconstruction unit 901 uses multi-source remote sensing data to fill in the missing data, obtaining more complete data, improving resolution, and completing the spatiotemporal sequence reconstruction to obtain the spectral feature time series of each pixel in the target area. Spatiotemporal sequence reconstruction includes the reconstruction of historical multi-source remote sensing data and the reconstruction of near-real-time multi-source remote sensing data to obtain near-real-time spectral feature time series and historical spectral feature time series from different periods.

[0210] The multi-dimensional feature fusion unit 902 is used to acquire the terrain and climate features of the target area and fuse them with the time series of spectral features to obtain the multi-dimensional feature vector of each pixel.

[0211] Specifically, by using various relevant data statistics websites, publicly available data on the topographic and climatic features of the target area at different times are obtained. The multidimensional feature fusion unit 902 then fuses these data with the spectral feature time series to obtain the multidimensional feature vector of each pixel.

[0212] The land cover classification calculation unit 903 is used to obtain the historical land cover classification of each pixel in the first historical reference period, and calculate the near real-time land cover classification of each pixel based on the historical land cover classification and multi-dimensional feature vector.

[0213] Specifically, an appropriate year can be selected as the first base year based on the actual situation. For example, the first historical base year can be selected with reference to the recommendations of the United Nations Convention to Combat Desertification (UNCCD). This application does not impose any restrictions on this.

[0214] The land degradation assessment unit 904 is used to calculate the land degradation assessment index of each pixel based on the multidimensional feature vector of each pixel and / or the historical land cover classification and the near real-time land cover classification, and to assess the land degradation status of the pixel.

[0215] Specifically, the multidimensional feature vector of each pixel contains the reflectance of various bands in the spectral feature time series, such as the reflectance of red light, blue light, and near-infrared bands, as well as various topographic and climatic features. Based on these parameters, multiple land assessment indicators for the target area can be calculated, and the land can be judged as to whether it has degraded in the corresponding aspects based on these indicators.

[0216] The near-real-time land degradation monitoring system provided in this application, through a time series reconstruction unit 901, a multi-dimensional feature fusion unit 902, a land cover classification calculation unit 903, and a land degradation assessment unit 904, achieves rapid prediction and utilization of land cover classification from near-real-time data, ensuring the real-time nature of monitoring results. It also fully utilizes various near-real-time data to calculate degradation indicators in the monitored area in multiple aspects such as land vegetation cover, land productivity, and soil carbon, realizing monitoring of multiple aspects of land degradation, forming a complete monitoring system, and enhancing the accuracy and completeness of monitoring results.

[0217] Among them, the multi-source remote sensing data includes high spatial and low temporal resolution remote sensing data and low spatial and high temporal resolution remote sensing data, and the spectral feature time series is a high spatial and high temporal resolution spectral feature time series.

[0218] Based on the above embodiments, further, such as Figure 10 As shown, the land degradation monitoring system provided in this application further includes: a multidimensional feature vector comprising historical multidimensional feature vectors and near-real-time multidimensional feature vectors; and a land cover classification calculation unit 903 comprising:

[0219] The change vector calculation module 903.1 is used to randomly select pixels as sample points and calculate the change vector of the sample points based on the historical multidimensional feature vector and the near real-time multidimensional feature vector of the sample points in the first historical reference period.

[0220] The sample classification migration module 903.2 is used to migrate the historical land cover classification of sample points whose angle and length of change vector are both less than a preset first threshold to the near real-time land cover classification of the corresponding sample points, thereby obtaining a migration sample set;

[0221] Classification model training module 903.3 is used to train a pre-built land cover classification model using a transfer sample set;

[0222] The land cover classification prediction module 903.4 is used to input the near real-time multidimensional feature vector of each pixel in the target area into the trained land cover classification model to obtain the near real-time land cover classification of each pixel.

[0223] The near-real-time land degradation monitoring system provided in this application obtains the near-real-time land cover classification for each pixel through the change vector calculation module 903.1, sample classification transfer module 903.2, classification model training module 903.3, and land cover classification prediction module 903.4. This enables rapid prediction of land cover classification based on near-real-time data, saves a significant amount of time in obtaining land cover labels through surveys and statistics, realizes the utilization of near-real-time data, and enhances the real-time nature of monitoring results.

[0224] Based on the above embodiments, further, such as Figure 11 As shown, the multidimensional feature vector includes: historical multidimensional feature vector and near-real-time multidimensional feature vector; the land degradation assessment indicators include historical values ​​of vegetation index and vegetation productivity index, as well as near-real-time values ​​of vegetation index and vegetation productivity index; the land degradation assessment unit 904 includes:

[0225] The historical index calculation module 904.1 is used to calculate the historical index values ​​of vegetation index and vegetation productivity index based on historical multidimensional feature vectors.

[0226] The near-real-time index calculation module 904.2 is used to calculate the near-real-time index values ​​of vegetation index and vegetation productivity index based on the near-real-time multidimensional feature vector.

[0227] The mean and variance calculation module 904.3 is used to calculate the mean and variance of each historical indicator value of vegetation index and vegetation productivity index;

[0228] The land degradation judgment module 904.4 is used to determine whether the mean of each near-real-time indicator value of vegetation index and vegetation productivity index differs from the mean of the corresponding historical indicator value by more than the variance. If so, the land is considered to be degraded.

[0229] The near-real-time land degradation monitoring system provided in this application, through historical index calculation module 904.1, near-real-time index calculation module 904.2, mean and variance calculation module 904.3, and land degradation judgment module 904.4, enables the detection of land vegetation health through multiple indicators and the judgment of whether the land has degraded in terms of vegetation health.

[0230] Based on the above embodiments, further, such as Figure 12 As shown, the land degradation assessment indicators include land cover change sub-indicators; land degradation assessment unit 904 includes:

[0231] The Land Cover Vegetation Degradation Assessment Module 904.5 is used to compare historical land cover classification with near-real-time land cover classification. It obtains land cover change sub-indicators for pixels based on a preset land cover transformation matrix. These sub-indicators represent the degradation status of land cover vegetation.

[0232] The near real-time land degradation monitoring system provided in this application obtains land cover change sub-indicators of pixels through the land cover vegetation degradation assessment module 904.5. The land cover change sub-indicators represent the degradation status of land cover vegetation, thereby realizing the monitoring of land vegetation cover status and assessing whether there is degradation in land vegetation cover.

[0233] exist Figure 11 Based on the embodiments, further, such as Figure 13 As shown, the land degradation monitoring system provided in this application also includes:

[0234] The regression analysis unit 905 is used to select any one of the vegetation index and vegetation productivity index as the first vegetation index, and to perform regression analysis on the historical index value and near real-time index value of the first vegetation index of each pixel to obtain the slope and significance value of the linear regression of each pixel.

[0235] The land degradation trend acquisition unit 906 is used to determine the land degradation trend of each pixel based on the slope, significance value and a preset second threshold.

[0236] The mean and standard deviation calculation unit 907 is used to select any one of the vegetation index and vegetation productivity index as the second vegetation index, calculate the mean and standard deviation of the historical index value of the second vegetation index of each pixel in the second historical reference period, and the mean of the historical index value and the near real-time index value of the second vegetation index of each pixel in the monitoring target period.

[0237] The significance calculation unit 908 is used to calculate the significance of the second vegetation index based on the mean and standard deviation of the second historical baseline period and the mean of the monitoring target period.

[0238] The land degradation status acquisition unit 909 is used to determine the land degradation status of each pixel based on the significance of the second vegetation index and the preset third threshold.

[0239] The land degradation performance value calculation unit 910 is used to select any one of the vegetation index and vegetation productivity index as the third vegetation index, divide the monitoring area into multiple target zones, and calculate the land degradation performance value of each pixel based on the maximum value of the near real-time index value of the third vegetation index of each pixel in each target zone and the near real-time index value of the third vegetation index of each pixel.

[0240] The land degradation performance acquisition unit 911 is used to determine the land degradation performance of each pixel based on the land degradation performance value and a preset fourth threshold.

[0241] The first land degradation judgment unit 912 is used to judge whether the land degradation trend, land degradation state and land degradation performance of each pixel is degraded. If the land degradation state and the land degradation performance are both degraded and / or the land degradation trend is degraded, then the land productivity change sub-index of the corresponding pixel is degraded; otherwise, the land productivity change sub-index of the corresponding pixel does not show degradation.

[0242] The near-real-time land degradation monitoring system provided in this application, through regression analysis unit 905, land degradation trend acquisition unit 906, mean and standard deviation calculation unit 907, significance calculation unit 908, land degradation status acquisition unit 909, land degradation performance value calculation unit 910, land degradation performance acquisition unit 911, and first land degradation judgment unit 912, obtains whether each pixel point has degraded in terms of "trend", "status" and "performance", and performs comprehensive analysis of the three indicators to further determine whether the pixel point has degraded in terms of land productivity, forming a relatively complete land degradation monitoring system, making full use of remote sensing monitoring data and intermediate data, and obtaining more professional land degradation monitoring results.

[0243] exist Figure 9 Based on the embodiments, further, such as Figure 14 As shown, the multidimensional feature vector includes: historical multidimensional feature vector and near-real-time multidimensional feature vector; the land degradation assessment index includes the land productivity change sub-indicator; the land degradation assessment unit 904 includes:

[0244] The historical index calculation module 904.6 is used to calculate the historical index values ​​of vegetation index and vegetation productivity index based on historical multidimensional feature vectors.

[0245] The near-real-time index calculation module 904.7 is used to calculate the near-real-time index values ​​of vegetation index and vegetation productivity index based on near-real-time multidimensional feature vectors.

[0246] The regression analysis module 904.8 is used to select any one of the vegetation index and vegetation productivity index as the first vegetation index, and to perform regression analysis on the historical index value and near real-time index value of the first vegetation index of each pixel to obtain the slope and significance value of the linear regression of each pixel.

[0247] The land degradation trend acquisition module 904.9 is used to determine the land degradation trend of each pixel based on the slope, significance value and preset fifth threshold.

[0248] The mean and standard deviation calculation module 904.10 is used to select any one of the vegetation index and vegetation productivity index as the second vegetation index, calculate the mean and standard deviation of the historical index value of the second vegetation index of each pixel in the second historical reference period, and the mean of the historical index value and the near real-time index value of the second vegetation index of each pixel in the monitoring target period.

[0249] The significance calculation module 904.11 is used to calculate the significance of the second vegetation index based on the mean and standard deviation of the second historical baseline period and the mean of the monitoring target period.

[0250] The land degradation status acquisition module 904.12 is used to determine the land degradation status of each pixel based on the significance of the second vegetation index and the preset sixth threshold.

[0251] The land degradation performance value calculation module 904.13 is used to select any one of the vegetation index and vegetation productivity index as the third vegetation index, divide the monitoring area into multiple target zones, and calculate the land degradation performance value of each pixel based on the maximum value of the near real-time index value of the third vegetation index of each pixel in each target zone and the near real-time index value of the third vegetation index of each pixel.

[0252] The land degradation performance acquisition module 904.14 is used to determine the land degradation performance of each pixel based on the land degradation performance value and the preset seventh threshold.

[0253] The land degradation judgment module 904.15 is used to determine whether the land degradation trend, land degradation state, and land degradation performance of each pixel are degraded. If the land degradation state and the land degradation performance are both degraded and / or the land degradation trend is degraded, then the land productivity change sub-index of the corresponding pixel is degraded; otherwise, the land productivity change sub-index of the corresponding pixel is not degraded.

[0254] The near-real-time land degradation monitoring system provided in this application, through historical index calculation module 904.6, near-real-time index calculation module 904.7, regression analysis module 904.8, land degradation trend acquisition module 904.9, mean and standard deviation calculation module 904.10, significance calculation module 904.11, land degradation status acquisition module 904.12, land degradation performance value calculation module 904.13, land degradation performance acquisition module 904.14, and land degradation judgment module 904.15, obtains whether each pixel point has degraded in terms of "trend", "status" and "performance", and performs comprehensive analysis of the three indicators to further determine whether the pixel point has degraded in terms of land productivity, forming a relatively complete land degradation monitoring system. It makes full use of remote sensing monitoring data and intermediate data to obtain more professional land degradation monitoring results.

[0255] Based on the above embodiments, further, such as Figure 15 As shown, the land degradation monitoring system provided in this application also includes:

[0256] The soil carbon change sub-index calculation unit 913 is used to obtain the near real-time soil carbon estimate of the pixel and the soil carbon estimate of the historical reference year, and calculate the soil carbon change sub-index of the pixel based on the near real-time soil carbon estimate of the pixel, the soil carbon estimate of the historical reference year and the difference between the current year and the historical reference year.

[0257] The soil carbon degradation judgment unit 914 is used to judge the soil carbon degradation status of a pixel based on the soil carbon change sub-indicator and the preset threshold range.

[0258] The near real-time land degradation monitoring system provided in this application, through the soil carbon change sub-index calculation unit 913 and the soil carbon degradation status judgment unit 914, combined with the above embodiments, further improves the land degradation assessment system, enriches the data sources and monitoring angles of land degradation, and realizes a complete system for fully monitoring land degradation from multiple perspectives.

[0259] Based on the above embodiments, further, such as Figure 16 As shown, the land degradation monitoring system provided in this application also includes:

[0260] The vegetation health index acquisition unit 915 is used to acquire the historical vegetation health index and near-real-time vegetation health index of pixels.

[0261] The Health Index Mean and Variance Calculation Unit 916 is used to calculate the mean and variance of each indicator value in the historical vegetation health index.

[0262] The second land degradation judgment unit 917 is used to determine whether the difference between the mean value of each indicator in the near-real-time vegetation health index and the mean value of the corresponding historical indicator exceeds the variance. If so, the land is considered to be degraded.

[0263] The near-real-time land degradation monitoring system provided in this application, through the vegetation health index acquisition unit 915, the health index mean and variance calculation unit 916, and the second land degradation judgment unit 917, realizes the monitoring of whether the land has degraded in terms of vegetation health, and further improves the near-real-time land degradation monitoring system provided in this application.

[0264] Figure 17 This is a schematic diagram of the physical structure of an electronic device provided in an embodiment of the present invention, as shown below. Figure 17As shown, the electronic device may include: a processor 1701, a communications interface 1702, a memory 1703, and a communication bus 1704, wherein the processor 1701, the communications interface 1702, and the memory 1703 communicate with each other through the communication bus 1704. Processor 1701 can call logical instructions in memory 1703 to execute the following methods: acquire historical and near-real-time multi-source remote sensing data of the target area and perform spatiotemporal sequence reconstruction to obtain the spectral feature time series of each pixel in the target area, including the near-real-time spectral feature time series and historical spectral feature time series of different periods; acquire the topographic and climatic features of the target area and fuse them with the spectral feature time series to obtain the multidimensional feature vector of each pixel; acquire the historical land cover classification of each pixel in the first historical reference period, and calculate the near-real-time land cover classification of each pixel based on the historical land cover classification and the multidimensional feature vector; calculate the land degradation assessment index of each pixel based on the multidimensional feature vector of each pixel, and / or the historical land cover classification and the near-real-time land cover classification, and assess the land degradation status of the pixel.

[0265] Furthermore, the logical instructions in the aforementioned memory 1703 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0266] This embodiment discloses a computer program product, which includes a computer program stored on a computer-readable storage medium. The computer program includes program instructions, and when the program instructions are executed by the computer, the computer can execute the methods provided in the above-described method embodiments, such as: acquiring historical and near-real-time multi-source remote sensing data of a target area and performing spatiotemporal sequence reconstruction to obtain a spectral feature time series of each pixel in the target area, the spectral feature time series including a near-real-time spectral feature time series and historical spectral feature time series of different periods; acquiring topographic and climatic features of the target area and fusing them with the spectral feature time series to obtain a multidimensional feature vector of each pixel; acquiring the historical land cover classification of each pixel in a first historical reference period, calculating the near-real-time land cover classification of each pixel based on the historical land cover classification and the multidimensional feature vector; calculating the land degradation assessment index of each pixel based on the multidimensional feature vector of each pixel, and / or the historical land cover classification and the near-real-time land cover classification, and assessing the land degradation status of the pixel.

[0267] This embodiment provides a computer-readable storage medium storing a computer program that causes the computer to execute the methods provided in the above-described method embodiments. For example, the methods include: acquiring historical and near-real-time multi-source remote sensing data of a target area and performing spatiotemporal sequence reconstruction to obtain a spectral feature time series for each pixel in the target area, the spectral feature time series including a near-real-time spectral feature time series and historical spectral feature time series from different periods; acquiring topographic and climatic features of the target area and fusing them with the spectral feature time series to obtain a multidimensional feature vector for each pixel; acquiring the historical land cover classification for each pixel in a first historical reference period, calculating the near-real-time land cover classification for each pixel based on the historical land cover classification and the multidimensional feature vector; calculating a land degradation assessment index for each pixel based on the multidimensional feature vector, and / or the historical land cover classification and the near-real-time land cover classification, and assessing the land degradation status of the pixels.

[0268] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0269] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0270] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0271] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0272] In the description of this specification, the references to terms such as "an embodiment," "a specific embodiment," "some embodiments," "for example," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0273] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A near-real-time land degradation monitoring method, characterized in that, include: Historical and near-real-time multi-source remote sensing data of the target area are acquired and spatiotemporal sequence reconstruction is performed to obtain the spectral feature time series of each pixel in the target area. The spectral feature time series includes the near-real-time spectral feature time series and the historical spectral feature time series of different periods. The terrain and climate features of the target area are obtained and fused with the time series of the spectral features to obtain a multidimensional feature vector for each pixel. Obtain the historical land cover classification of each pixel in the first historical reference period, and calculate the near real-time land cover classification of each pixel based on the historical land cover classification and the multi-dimensional feature vector. Based on the multidimensional feature vector of each pixel, and / or historical land cover classification and near real-time land cover classification, calculate the land degradation assessment index of each pixel, and assess the land degradation status of the pixel; The multidimensional feature vector includes: a historical multidimensional feature vector and a near-real-time multidimensional feature vector; the calculation of the near-real-time land cover classification for each pixel based on the historical land cover classification and the multidimensional feature vector includes: Randomly select pixels as sample points, and calculate the change vector of the sample points based on the historical multidimensional feature vector and the near real-time multidimensional feature vector of the sample points in the first historical reference period. The historical land cover classification of sample points whose angle and length of the change vector are both less than a preset first threshold are migrated to the near real-time land cover classification of the corresponding sample points to obtain a migration sample set. The pre-established land cover classification model was trained using the aforementioned migration sample set; The near real-time multidimensional feature vectors of each pixel in the target region are input into the trained land cover classification model to obtain the near real-time land cover classification of each pixel. The land degradation assessment index includes a sub-index of land productivity change; the calculation of the land degradation assessment index for each pixel based on the multidimensional feature vector of each pixel, and / or historical land cover classification and near-real-time land cover classification, and the assessment of the land degradation status of the pixels, includes: The historical index values ​​of vegetation index and vegetation productivity index are calculated based on the historical multidimensional feature vector. The near-real-time index values ​​of the vegetation index and vegetation productivity index are calculated based on the near-real-time multidimensional feature vector. Select any one of the vegetation index and vegetation productivity index as the first vegetation index, and perform regression analysis on the historical index value and near real-time index value of the first vegetation index of each pixel to obtain the slope and significance value of the linear regression of each pixel. The land degradation trend of each pixel is determined based on the slope, significance value, and preset fifth threshold. Select any one of the vegetation index and vegetation productivity index as the second vegetation index, calculate the mean and standard deviation of the historical index value of the second vegetation index of each pixel in the second historical reference period, and the mean of the historical index value and the near real-time index value of the second vegetation index of each pixel in the monitoring target period. The significance of the second vegetation index is calculated based on the mean and standard deviation of the second historical baseline period and the mean of the monitoring target period; The land degradation status of each pixel is determined based on the significance of the second vegetation index and the preset sixth threshold. Select any one of the vegetation index and vegetation productivity index as the third vegetation index, divide the monitoring area into multiple target zones, and calculate the land degradation performance value of each pixel based on the maximum value of the near real-time index value of the third vegetation index of each pixel in each target zone and the near real-time index value of the third vegetation index of each pixel. The land degradation performance of each pixel is determined based on the land degradation performance value and the preset seventh threshold. Determine whether the land degradation trend, land degradation state, and land degradation performance of each pixel are degraded. If both the land degradation state and the land degradation performance are degraded and / or the land degradation trend is degraded, then the land productivity change sub-index of the corresponding pixel is degraded; otherwise, the land productivity change sub-index of the corresponding pixel is not degraded.

2. The near-real-time land degradation monitoring method according to claim 1, characterized in that, The multi-source remote sensing data includes high spatial and low temporal resolution remote sensing data and low spatial and high temporal resolution remote sensing data, and the spectral feature time series is a high spatial and high temporal resolution spectral feature time series.

3. The near-real-time land degradation monitoring method according to claim 1, characterized in that, The multidimensional feature vector includes: historical multidimensional feature vector and near-real-time multidimensional feature vector; the land degradation assessment index includes historical values ​​of vegetation index and vegetation productivity index, and near-real-time values ​​of vegetation index and vegetation productivity index; the calculation of land degradation assessment index for each pixel based on the multidimensional feature vector of each pixel, and / or historical land cover classification and near-real-time land cover classification, and the assessment of the land degradation status of the pixel, includes: Calculate the historical index values ​​of the vegetation index and vegetation productivity index based on the historical multidimensional feature vector; The near-real-time index values ​​of the vegetation index and vegetation productivity index are calculated based on the near-real-time multidimensional feature vector. Calculate the mean and variance of each historical value of the vegetation index and vegetation productivity index; Determine whether the mean of each near-real-time indicator value of the vegetation index and vegetation productivity index differs from the mean of the corresponding historical indicator values ​​by more than the variance. If so, the land is considered to be degraded.

4. The near-real-time land degradation monitoring method according to claim 1, characterized in that, The land degradation assessment index includes a land cover change sub-index; the calculation of the land degradation assessment index for each pixel based on the multidimensional feature vector of each pixel, and / or historical land cover classification and near-real-time land cover classification, and the assessment of the land degradation status of the pixels, includes: By comparing the historical land cover classification with the near real-time land cover classification, the land cover change sub-index of the pixel is obtained according to the preset land cover transformation matrix. The land cover change sub-index represents the degradation of land cover vegetation.

5. The near-real-time land degradation monitoring method according to claim 3, characterized in that, Also includes: Select any one of the vegetation index and vegetation productivity index as the first vegetation index, and perform regression analysis on the historical index value and near real-time index value of the first vegetation index of each pixel to obtain the slope and significance value of the linear regression of each pixel. The land degradation trend of each pixel is determined based on the slope, significance value, and preset second threshold. Select any one of the vegetation index and vegetation productivity index as the second vegetation index, calculate the mean and standard deviation of the historical index value of the second vegetation index of each pixel in the second historical reference period, and the mean of the historical index value and the near real-time index value of the second vegetation index of each pixel in the monitoring target period. The significance of the second vegetation index is calculated based on the mean and standard deviation of the second historical baseline period and the mean of the monitoring target period; The land degradation status of each pixel is determined based on the significance of the second vegetation index and a preset third threshold. Select any one of the vegetation index and vegetation productivity index as the third vegetation index, divide the monitoring area into multiple target zones, and calculate the land degradation performance value of each pixel based on the maximum value of the near real-time index value of the third vegetation index of each pixel in each target zone and the near real-time index value of the third vegetation index of each pixel. The land degradation performance of each pixel is determined based on the land degradation performance value and the preset fourth threshold. Determine whether the land degradation trend, land degradation state, and land degradation performance of each pixel are degraded. If both the land degradation state and the land degradation performance are degraded and / or the land degradation trend is degraded, then the land productivity change sub-index of the corresponding pixel is degraded; otherwise, the land productivity change sub-index of the corresponding pixel is not degraded.

6. The near-real-time land degradation monitoring method according to claim 1, characterized in that, Also includes: Obtain the near-real-time soil carbon estimate of the pixel and the soil carbon estimate of the historical baseline year. Calculate the soil carbon change sub-index of the pixel based on the near-real-time soil carbon estimate of the pixel, the soil carbon estimate of the historical baseline year, and the difference between the current year and the historical baseline year. The soil carbon degradation status of the pixel is determined based on the soil carbon change sub-indicator and the preset threshold range.

7. The near-real-time land degradation monitoring method according to claim 1, characterized in that, Also includes: Obtain the historical and near-real-time vegetation health indices of pixels; Calculate the mean and variance of each indicator value in the historical vegetation health index; Determine whether the difference between the mean value of each indicator in the near-real-time vegetation health index and the mean value of the corresponding historical indicator exceeds the variance. If so, the land is considered to be degraded.

8. A near real-time land degradation monitoring system, characterized in that, include: The time series reconstruction unit is used to acquire historical and near-real-time multi-source remote sensing data of the target area and perform spatiotemporal sequence reconstruction to obtain the spectral feature time series of each pixel in the target area. The spectral feature time series includes the near-real-time spectral feature time series and the historical spectral feature time series of different periods. A multi-dimensional feature fusion unit is used to acquire the terrain and climate features of the target area and fuse them with the time series of the spectral features to obtain the multi-dimensional feature vector of each pixel. The land cover classification calculation unit is used to obtain the historical land cover classification of each pixel in the first historical reference period, and to calculate the near real-time land cover classification of each pixel based on the historical land cover classification and the multi-dimensional feature vector. The land degradation assessment unit is used to calculate the land degradation assessment index of each pixel based on the multidimensional feature vector of each pixel and / or historical land cover classification and near real-time land cover classification, and to assess the land degradation status of the pixel. The multidimensional feature vector includes: historical multidimensional feature vector and near real-time multidimensional feature vector; the land cover classification calculation unit includes: The change vector calculation module is used to randomly select pixels as sample points and calculate the change vector of the sample points based on the historical multidimensional feature vector and the near real-time multidimensional feature vector of the sample points in the first historical reference period. The sample classification migration module is used to migrate the historical land cover classification of sample points whose angle and length of the change vector are both less than a preset first threshold to the near real-time land cover classification of the corresponding sample points, thereby obtaining a migration sample set. A classification model training module is used to train a pre-established land cover classification model using the migration sample set; The land cover classification prediction module is used to input the near real-time multidimensional feature vector of each pixel in the target area into the trained land cover classification model to obtain the near real-time land cover classification of each pixel. The land degradation assessment indicators include sub-indicators of land productivity change; the land degradation assessment unit includes: The historical index calculation module is used to calculate the historical index values ​​of vegetation index and vegetation productivity index based on the historical multidimensional feature vector. The near-real-time indicator calculation module is used to calculate the near-real-time indicator values ​​of the vegetation index and the vegetation productivity index based on the near-real-time multidimensional feature vector. The regression analysis module is used to select any one of the vegetation index and vegetation productivity index as the first vegetation index, and to perform regression analysis on the historical index value and near real-time index value of the first vegetation index of each pixel to obtain the slope and significance value of the linear regression of each pixel. The land degradation trend acquisition module is used to determine the land degradation trend of each pixel based on the slope, significance value, and a preset fifth threshold. The mean and standard deviation calculation module is used to select any one of the vegetation index and vegetation productivity index as the second vegetation index, calculate the mean and standard deviation of the historical index value of the second vegetation index of each pixel in the second historical reference period, and the mean of the historical index value and the near real-time index value of the second vegetation index of each pixel in the monitoring target period. The significance calculation module is used to calculate the significance of the second vegetation index based on the mean and standard deviation of the second historical baseline period and the mean of the monitoring target period; The land degradation status acquisition module is used to determine the land degradation status of each pixel based on the significance of the second vegetation index and the preset sixth threshold. The land degradation performance value calculation module is used to select any one of the vegetation index and vegetation productivity index as the third vegetation index, divide the monitoring area into multiple target zones, and calculate the land degradation performance value of each pixel based on the maximum value of the near real-time index value of the third vegetation index of each pixel in each target zone and the near real-time index value of the third vegetation index of each pixel. The land degradation performance acquisition module is used to determine the land degradation performance of each pixel based on the land degradation performance value and a preset seventh threshold. The land degradation judgment module is used to determine whether the land degradation trend, the land degradation state, and the land degradation performance of each pixel are degraded. If the land degradation state and the land degradation performance are both degraded and / or the land degradation trend is degraded, then the land productivity change sub-index of the corresponding pixel is degraded; otherwise, the land productivity change sub-index of the corresponding pixel is not degraded.

9. The near-real-time land degradation monitoring system according to claim 8, characterized in that, The multidimensional feature vector includes: historical multidimensional feature vector and near-real-time multidimensional feature vector; the land degradation assessment indicators include historical values ​​of vegetation index and vegetation productivity index, and near-real-time values ​​of vegetation index and vegetation productivity index; the land degradation assessment unit includes: The historical index calculation module is used to calculate the historical index values ​​of the vegetation index and the vegetation productivity index based on the historical multidimensional feature vector. The near-real-time indicator calculation module is used to calculate the near-real-time indicator values ​​of the vegetation index and the vegetation productivity index based on the near-real-time multidimensional feature vector. The mean and variance calculation module is used to calculate the mean and variance of each historical indicator value of the vegetation index and the vegetation productivity index. The land degradation judgment module is used to determine whether the difference between the mean of each near-real-time indicator value of the vegetation index and the vegetation productivity index and the mean of the corresponding historical indicator value exceeds the variance. If so, the land is considered to be degraded.

10. The near-real-time land degradation monitoring system according to claim 8, characterized in that, The land degradation assessment indicators include land cover change sub-indicators; the land degradation assessment unit includes: The land cover vegetation degradation assessment module is used to compare the historical land cover classification with the near real-time land cover classification, and obtain the land cover change sub-indices of the pixels according to the preset land cover transformation matrix. The land cover change sub-indices represent the degradation status of land cover vegetation.

11. The near-real-time land degradation monitoring system according to claim 9, characterized in that, Also includes: The regression analysis unit is used to select any one of the vegetation index and vegetation productivity index as the first vegetation index, and to perform regression analysis on the historical index value and near real-time index value of the first vegetation index of the pixel to obtain the slope and significance value of the linear regression of each pixel. The land degradation trend acquisition unit is used to determine the land degradation trend of each pixel based on the slope, significance value, and a preset second threshold. The mean and standard deviation calculation unit is used to select any one of the vegetation index and vegetation productivity index as the second vegetation index, calculate the mean and standard deviation of the historical index value of the second vegetation index of each pixel in the second historical reference period, and the mean of the historical index value and the near real-time index value of the second vegetation index of each pixel in the monitoring target period. A significance calculation unit is used to calculate the significance of the second vegetation index based on the mean and standard deviation of the second historical reference period and the mean of the monitoring target period; The land degradation status acquisition unit is used to determine the land degradation status of each pixel based on the significance of the second vegetation index and a preset third threshold. The land degradation performance value calculation unit is used to select any one of the vegetation index and vegetation productivity index as the third vegetation index, divide the monitoring area into multiple target zones, and calculate the land degradation performance value of each pixel based on the maximum value of the near real-time index value of the third vegetation index of each pixel in each target zone and the near real-time index value of the third vegetation index of each pixel. The land degradation performance acquisition unit is used to determine the land degradation performance of each pixel based on the land degradation performance value and a preset fourth threshold. The first land degradation judgment unit is used to determine whether the land degradation trend, the land degradation state, and the land degradation performance of each pixel are degraded. If the land degradation state and the land degradation performance are both degraded and / or the land degradation trend is degraded, then the land productivity change sub-index of the corresponding pixel is degraded; otherwise, the land productivity change sub-index of the corresponding pixel is not degraded.

12. The near-real-time land degradation monitoring system according to claim 8, characterized in that, Also includes: The soil carbon change sub-index calculation unit is used to obtain the near real-time soil carbon estimate of the pixel and the soil carbon estimate of the historical reference year, and calculate the soil carbon change sub-index of the pixel based on the near real-time soil carbon estimate of the pixel, the soil carbon estimate of the historical reference year and the difference between the current year and the historical reference year. The soil carbon degradation determination unit is used to determine the soil carbon degradation status of the pixel based on the soil carbon change sub-indicator and a preset threshold range.

13. The near-real-time land degradation monitoring system according to claim 8, characterized in that, Also includes: The vegetation health index acquisition unit is used to acquire the historical vegetation health index and near real-time vegetation health index of pixels. The health index mean and variance calculation unit is used to calculate the mean and variance of each indicator value in the historical vegetation health index. The second land degradation judgment unit is used to determine whether the difference between the mean value of each indicator in the near-real-time vegetation health index and the mean value of the corresponding historical indicator exceeds the variance. If so, the land is considered to be degraded.

14. A computer 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, it implements the method according to any one of claims 1 to 7.

15. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method according to any one of claims 1 to 7.

16. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the method according to any one of claims 1 to 7.

Images