Grassland degradation degree monitoring model and method based on remote sensing technology

Abstract

The invention discloses a grassland degradation degree monitoring model and method based on a remote sensing technology, and belongs to the technical field of ecological environment monitoring. The method comprises the following steps: step 1, constructing a multi-scale remote sensing data set with consistent time and space, and establishing a high-precision truth value library of grassland types and degradation degrees in combination with ground measured data; step 2, constructing a remote sensing index system adapted to different spatial scales; step 3, identifying key scale conversion nodes of grassland degradation driven by climate and human factors; step 4, grassland degradation degree grading from the pixel scale to the landscape scale is realized, and spatial distribution, area change and time-space evolution trend of each degradation grade are quantitatively analyzed; step 5, realizing reliable conversion and up-down transmission of analysis results of different scales, and generating a comprehensive grassland degradation monitoring result; and step 6, automatically generating differentiated management strategies, and outputting a grassland degradation level distribution map, a driving force analysis map, a risk early warning map and a partition management suggestion map.

Description

Technical Field

[0001] This application relates to the technical field of ecological environment monitoring. More specifically, it relates to a monitoring model and method for grassland degradation degree based on remote sensing technology. Background Art

[0002] Grassland degradation refers to the phenomenon that the grassland ecosystem is affected by natural factors and human activities, resulting in a decrease in vegetation coverage, deterioration of soil quality, and decline of ecological functions. Grassland degradation not only affects the stability of the ecological environment, but also leads to soil erosion, reduction of biodiversity, and decline of agricultural and pastoral productivity. With the increasingly serious problem of grassland degradation, how to achieve real-time monitoring and accurate assessment of the grassland degradation process has become a key problem that needs to be solved urgently worldwide.

[0003] Traditional grassland degradation monitoring methods usually rely on ground surveys and manual sampling. These methods are not only time-consuming and laborious, but also have certain limitations. It is difficult to cover large areas and cannot reflect the dynamic changes of grassland degradation in real time. With the rapid development of remote sensing technology, the remote sensing-based grassland monitoring method has gradually become the main means of grassland degradation monitoring because it can quickly obtain high-efficiency data of large areas. However, the existing remote sensing monitoring methods still face multiple challenges, including how to accurately evaluate the grassland degradation degree at multiple scales and multiple time scales, how to integrate remote sensing data from different sources, how to effectively identify the different impacts of climate and human factors on grassland degradation, and how to provide precise and operable support for management decisions.

[0004] Although some progress has been made in grassland degradation monitoring by remote sensing technology, the existing methods mostly stay at the monitoring of a single scale or a single indicator, lacking a comprehensive and integrated analysis of multi-dimensional characteristics. Especially in the aspect of multi-scale data fusion and degradation driving force analysis with spatio-temporal consistency, there are still technical gaps. At the same time, in the application level of the existing grassland degradation monitoring results, there are often no scientific and reasonable management strategies and measures, making it difficult to truly promote grassland protection and restoration work.

[0005] In summary, how to achieve precise monitoring, comprehensive analysis, and generation of an efficient management strategy for grassland degradation through remote sensing technology has become a technical problem that needs to be solved urgently. Summary of the Invention

[0006] In order to overcome a series of defects existing in the prior art, the purpose of this application is to provide a method for monitoring the grassland degradation degree based on remote sensing technology for the above problems, including the following steps:

[0007] Step 1, construct a spatio-temporally consistent multi-scale remote sensing data set, and combine it with ground measured data to establish a high-precision truth database of grassland types and degradation degrees;

[0008] Step 2: Automatically screen the most sensitive remote sensing indicators for grassland degradation and construct a remote sensing indicator system suitable for different spatial scales;

[0009] Step 3: Identify the key scale conversion nodes where grassland degradation is driven by climate and human factors;

[0010] Step 4: Realize the classification of grassland degradation degree from pixel scale to landscape scale, and quantitatively analyze the spatial distribution, area change and spatio-temporal evolution trend of each degradation level;

[0011] Step 5: Realize the reliable conversion and up-down transmission of analysis results at different scales, and generate comprehensive grassland degradation monitoring results;

[0012] Step 6: Automatically generate differentiated management strategies and output the grassland degradation grade distribution map, driving force analysis map, risk warning map and zoning management suggestion map.

[0013] Furthermore, Step 2 includes the following steps:

[0014] Based on the corrected multi-source remote sensing data, calculate the vegetation index, water index and soil index, and construct a comprehensive index set that can comprehensively characterize the structural and functional characteristics of the grassland ecosystem;

[0015] For each index, calculate its statistical characteristics in the time dimension and space dimension, construct a multi-dimensional feature space, and accurately depict the dynamic changes of the grassland ecosystem at different degradation stages;

[0016] At the pixel, patch and landscape three spatial scales, calculate the information gain value and mutual information value of each index for grassland degradation classification, quantitatively evaluate the classification contribution degree of each index at different scales, and screen out the candidate index set with significant and stable contributions through information stability analysis;

[0017] Evaluate the response sensitivity of each index to the grassland degradation gradient, calculate the sensitivity coefficient and discrimination index, and combined with the true value library data, analyze the performance differences of the index in different ecological regions, grassland types and degradation degree intervals, and identify the most discriminative characteristic index at each scale;

[0018] Automatically screen the optimal index combination at different spatial scales, ensure the classification accuracy while minimizing the redundancy between indexes, and comprehensively consider the economy and practicability of data acquisition to form an adaptive index combination scheme suitable for different scales and grassland types;

[0019] Establish cross-scale index transfer relationships and conversion rules, evaluate the classification performance and stability of each scale index system based on an independent verification dataset, and finally construct a grassland monitoring index system with scale self-adaptability, regional pertinence and degradation sensitivity.

[0020] Furthermore, for each indicator, calculate its statistical characteristics in the time dimension and the spatial dimension, construct a multi-dimensional feature space, and accurately characterize the dynamic changes of the grassland ecosystem at different degradation stages, including the following steps:

[0021] For each indicator, calculate the time mean μ t , the time variance σ t 2 , the change rate V t and the autocorrelation R t to capture the stability, volatility and trend information of the grassland ecosystem at different times;

[0022] At each moment, calculate the spatial mean μ s , the spatial variance σ s 2 and the spatial autocorrelation I s for all indicator data to quantify the balance, discreteness and adjacent relationship of the ecosystem in spatial distribution;

[0023] Combine the statistical characteristics extracted from the time and spatial dimensions into a multi-dimensional feature vector F = [μ t , σ t 2 , V t , R t , μ s , σ s 2 , I s to form a feature space that comprehensively characterizes the dynamic changes of grassland degradation;

[0024] By comparing the feature vectors of different periods or regions and calculating the ecosystem dynamic change index D, achieve accurate determination and dynamic monitoring of the degradation stage of the grassland ecosystem. The formula is expressed as: where μ t (t) is the indicator mean in the time dimension, representing the mean of the grassland degradation indicator calculated at time point t; μ s (s) is the indicator mean in the spatial dimension, the mean of the grassland degradation indicator calculated in spatial unit s; μ t (t + 1) is the indicator mean in the time dimension, the mean of the grassland degradation indicator calculated at time point t + 1; μ s (s + 1) is the indicator mean in the spatial dimension, the mean of the grassland degradation indicator calculated in spatial unit s + 1.

[0025] Furthermore, at the three spatial scales of pixel, patch and landscape, calculate the information gain value and mutual information value of each indicator for grassland degradation classification, quantitatively evaluate the classification contribution degree of each indicator at different scales, and screen out a candidate indicator set with significant and stable contributions through information stability analysis, including the following steps:

[0026] Based on the remote sensing data of the target area, three spatial scales of pixels, patches, and landscapes are divided;

[0027] Under each spatial scale, collect the corresponding grassland degradation classification labels and various remote sensing index data to ensure that the dataset under each scale contains sufficient sample size and spatial information;

[0028] For each spatial scale, use the information gain formula IG(T,A) = H(T) - H(T|A) to calculate the information gain value of each remote sensing index A for the grassland degradation classification T. Among them, IG(T,A) represents the information gain of the remote sensing index A for the grassland degradation classification T, that is, the degree of reduction in the uncertainty of the grassland degradation classification under the condition of knowing A; H(T) represents the entropy of the grassland degradation classification T, measuring the overall uncertainty of T; H(T|A) represents the conditional entropy of the grassland degradation classification T after the given remote sensing index A, that is, the remaining uncertainty of T under the condition of knowing A;

[0029] Use the mutual information formula MI(A,T) = H(T) + H(A) - H(T,A) to calculate the mutual information value between the remote sensing index and the grassland degradation classification at different spatial scales. Among them, MI(A,T) represents the mutual information between the remote sensing index A and the grassland degradation classification T; H(A) represents the entropy of the remote sensing index A; H(T,A) represents the joint entropy of the grassland degradation classification T and the remote sensing index A;

[0030] Quantitatively evaluate the classification contribution degree of each remote sensing index at different spatial scales through the information gain value and the mutual information value;

[0031] Calculate the standard deviation of the information gain or mutual information value of the remote sensing index at different spatial scales, and use the formula: Evaluate the stability of the information, where μ IG represents the mean value of the information gain values calculated at different spatial scales, used to evaluate the overall classification contribution degree of the index; σ IG represents the standard deviation of the information gain values calculated at different spatial scales, used to measure the stability of the classification contribution degree of this index at each scale;

[0032] According to the results of the information stability analysis, screen out the remote sensing index set that is stable and has significant contributions at different scales.

[0033] Furthermore, step 3 includes the following steps:

[0034] Integrate climate factor and human factor data to establish a spatio-temporally consistent potential driving force database for grassland degradation, ensuring that the data of each driving factor matches the remote sensing monitoring data in terms of time and space resolution;

[0035] Taking the grassland degradation index as the response variable and climate and anthropogenic factors as the predictive variables, a multi-level Bayesian network model was constructed, where: the first level represents the pixel-scale response, the second level represents the patch-scale response, and the third level represents the landscape-scale response. The levels are connected through conditional probability relationships;

[0036] Calculate the main effects, interaction effects of climate and anthropogenic factors and their contribution rates at different scales, and use partial correlation analysis and variance decomposition techniques to separate the independent contributions of each factor;

[0037] Combined with information criteria and Bayesian factor comparison, identify the spatial scale conversion threshold of the dominant effects of climate and anthropogenic factors, and use the entropy maximization criterion to determine the optimal scale segmentation point, construct a scale-driving force response curve to identify key scale nodes;

[0038] Calculate the relative contribution rates of each driving factor to grassland degradation and their spatial differentiation characteristics at different scales, generate a scale-driving force contribution rate map, and achieve precise quantification and spatial visualization of the impacts of climate and anthropogenic factors.

[0039] Furthermore, calculating the main effects, interaction effects of climate and anthropogenic factors and their contribution rates at different scales, and using partial correlation analysis and variance decomposition techniques to separate the independent contributions of each factor includes the following steps:

[0040] Use a multiple linear regression model, set the grassland degradation degree as the dependent variable, and climate factors and anthropogenic factors as independent variables. Obtain the main effect coefficients through regression analysis and calculate the contribution rate of each main effect to the total variance;

[0041] Add the interaction term of climate factors and anthropogenic factors to the regression model, calculate the interaction effect coefficient, and calculate the contribution rate of the interaction effect through analysis of variance to measure the impact of the interaction between climate and anthropogenic factors on grassland degradation;

[0042] By calculating the partial correlation coefficient and respectively evaluate the independent contributions of climate and anthropogenic factors to grassland degradation under the condition of controlling the influence of another factor. Among them, represents the partial correlation coefficient between variable X1 and the dependent variable Y after controlling variable X2; represents the partial correlation coefficient between variable X2 and the dependent variable Y after controlling variable X1;

[0043] Apply the variance decomposition method to decompose the total variance into the variances of climate factors, anthropogenic factors, interaction effects, and error terms;

[0044] By calculating the variance contribution rate of each part, clarify the contribution degree of each factor to the variation of grassland degradation.

[0045] Further, step 4 includes the following steps:

[0046] Extract local texture features using an improved convolutional neural network at the pixel scale, integrate an attention mechanism at the patch scale to enhance the feature expression of key regions, and introduce a graph convolutional network at the landscape scale to capture the topological relationships between landscape units;

[0047] Utilize spatial context information to optimize the classification boundary and achieve high-precision identification of grassland degradation levels at a resolution of 1–30 meters;

[0048] Perform multi-scale segmentation on high-resolution remote sensing images to generate homogeneous patches with ecological significance, achieve degradation classification at the grassland functional unit level, and analyze the impacts of grassland fragmentation and connectivity changes on ecological functions;

[0049] Capture the macroscopic patterns and functional changes of the grassland ecosystem, analyze the spatial aggregation and diffusion trends of degraded areas, and evaluate the integrity of the ecosystem;

[0050] Based on the classification results of multiple periods, construct a degradation transfer matrix to quantitatively calculate the area proportions, spatial distribution characteristics, and temporal evolution laws of each degradation level at different scales.

[0051] Further, step 5 includes the following steps:

[0052] Establish a conversion rule library between the pixel, patch, and landscape scales. Specifically: for discrete classification results, use the dominant type method and weighted voting method for upscaling aggregation; for continuous degradation indices, use the regional average method and area weighting method for scale conversion; meanwhile, design a non-linear conversion function considering spatial heterogeneity to capture the mutation effect of degradation characteristics during scale conversion;

[0053] Use the global information at the high-level scale as prior knowledge to constrain the local detail expression at the lower-level scale, and at the same time use the high-precision information at the lower-level scale as evidence to update the overall cognition at the upper-level scale, constructing a two-way information flow mechanism;

[0054] Calculate the entropy value, variance, and confidence interval of each spatial unit at different scales, identify high-uncertainty regions, and analyze the main factors leading to uncertainty;

[0055] Identify the grassland types and degradation processes that are highly sensitive to a specific scale, calculate the scale adaptability index, quantitatively evaluate the applicability of the monitoring results at each scale, and construct an optimal scale selection decision tree;

[0056] Fuse the monitoring results at each scale according to the reliability weights to generate a comprehensive grassland degradation monitoring result with a unified accuracy standard and complete spatial coverage.

[0057] Further, step 6 includes the following steps:

[0058] Based on the grassland degradation level classification results and spatial statistical analysis, generate a high-precision degradation level distribution map, and visually display the spatial distribution characteristics of areas with different degradation degrees through a graded color scale;

[0059] Construct a spatial contribution rate model of climate factors and human factors, generate a driving force analysis map, and use a two-factor color coding method to visually present the spatial heterogeneity of the dominant factors;

[0060] Combine historical data sequences with current monitoring results, generate a grassland degradation risk warning map, identify potential rapid degradation areas and ecological fragile points, and set corresponding warning thresholds according to the risk levels;

[0061] According to the grassland degradation degree, driving factors, and ecological sensitivity, divide management units, generate a zoning management advice map, and propose targeted management measures to achieve the efficiency and pertinence of grassland protection and restoration.

[0062] The grassland degradation degree monitoring model based on remote sensing technology includes the following modules:

[0063] Data acquisition and preprocessing module, responsible for integrating multi-source remote sensing data and ground measured data, constructing a spatio-temporally consistent multi-scale data set, and performing preprocessing to ensure the accuracy and consistency of the data;

[0064] Degradation degree evaluation module, based on the preprocessed remote sensing data, calculate relevant indicators and perform multi-dimensional feature analysis, and screen out the optimal indicator combination;

[0065] Driving force analysis module, quantify the main effects, interaction effects and their contribution rates of climate and human factors on grassland degradation, and identify key scale nodes;

[0066] Result display and management module, display the grassland degradation level distribution, driving force analysis, risk warning and management advice in an intuitive chart form, and perform management;

[0067] Multi-scale analysis and conversion module, establish conversion rules between different spatial scales, realize multi-scale segmentation and feature extraction, construct a two-way information flow mechanism, and ensure the reliable conversion and fusion of analysis results at different scales;

[0068] Management strategy generation module, automatically generate differentiated management strategies according to the grassland degradation monitoring results and related analyses, propose targeted management measures, and achieve the efficiency and pertinence of grassland protection and restoration.

[0069] Compared with the prior art, the present application has the following beneficial effects:

[0070] This application realizes the dynamic monitoring of grassland degradation degree, the analysis of degradation driving forces, and the generation of differentiated management strategies by constructing a multi-scale remote sensing dataset, comprehensively analyzing climate and anthropogenic factors, and combining with a high-precision ground truth database, providing an accurate, visual, and operable grassland degradation assessment and management solution. Description of the Drawings

[0071] Figure 1 It is a schematic flowchart of the method for monitoring grassland degradation degree based on remote sensing technology disclosed in the embodiments of this application.

[0072] Figure 2 It is a schematic structural diagram of the monitoring model for grassland degradation degree based on remote sensing technology disclosed in the embodiments of this application. Detailed Embodiments

[0073] To make the objectives, technical solutions, and advantages of the implementation of the present invention clearer, the technical solutions in the embodiments of the present invention will be described in more detail below with reference to the accompanying drawings in the embodiments of the present invention. In the drawings, the same or similar reference numerals denote the same or similar elements or elements with the same or similar functions throughout. The described embodiments are some, but not all, of the embodiments of the present invention.

[0074] All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0075] The embodiments described below with reference to the drawings and directional terms are exemplary and are intended to explain the present invention and should not be construed as limiting the present invention.

[0076] As Figure 1 shown, the method for monitoring grassland degradation degree based on remote sensing technology includes the following steps:

[0077] Step 1: Construct a spatio-temporally consistent multi-scale remote sensing dataset, and establish a high-precision ground truth database for grassland types and degradation degrees in combination with ground measured data;

[0078] Step 2: Automatically screen the remote sensing indicators most sensitive to grassland degradation, and construct a remote sensing indicator system suitable for different spatial scales;

[0079] Step 3: Identify the key scale conversion nodes where grassland degradation is driven by climate and anthropogenic factors;

[0080] Step 4: Realize the classification of grassland degradation degree from the pixel scale to the landscape scale, and quantitatively analyze the spatial distribution, area change, and spatio-temporal evolution trend of each degradation level;

[0081] Step 5: Achieve reliable conversion and up-down transfer of analysis results at different scales to generate comprehensive grassland degradation monitoring results;

[0082] Step 6: Automatically generate differentiated management strategies and output grassland degradation level distribution maps, driving force analysis maps, risk early warning maps, and regional management suggestion maps.

[0083] In summary, the grassland degradation degree monitoring method based on remote sensing technology effectively improves the accuracy and spatio-temporal adaptability of grassland degradation monitoring through a series of precise data processing and analysis steps. First, a spatio-temporally consistent multi-scale remote sensing data set is constructed, and combined with ground measured data, a high-precision grassland type and degradation degree truth database is established, providing reliable basic data for subsequent analysis. Then, the remote sensing indicators that can best reflect grassland degradation are automatically selected, and a remote sensing indicator system suitable for different spatial scales is constructed, which ensures the multi-dimensionality and accuracy of the monitoring results. When identifying the key driving factors of grassland degradation, this method particularly focuses on the key scale conversion nodes under the action of climate and human factors, thus revealing the complex spatio-temporal dynamics of grassland degradation. Next, through the classification of grassland degradation degree from pixel scale to landscape scale, the precise definition of degradation levels is achieved, and by quantitatively analyzing its spatial distribution, area change, and spatio-temporal evolution trend, the depth and breadth of the monitoring results are further enhanced. In addition, this method realizes reliable conversion and up-down transfer of analysis results between different scales, enabling effective communication of monitoring data between different spatial levels, and finally generating comprehensive grassland degradation monitoring results. On this basis, this method can automatically generate differentiated grassland management strategies, and by outputting grassland degradation level distribution maps, driving force analysis maps, risk early warning maps, and regional management suggestion maps, it provides practical decision-making support for grassland protection and restoration.

[0084] Furthermore, Step 1 includes the following steps:

[0085] Obtain high-resolution optical satellite images for refined grassland mapping, medium and low-resolution optical satellite images for long-term dynamic monitoring, and microwave remote sensing data for observing grassland biomass and moisture conditions through clouds, and perform strip removal, broken line repair, and denoising processing on all images;

[0086] Based on a high-precision control point network, perform geometric precise correction on remote sensing images from different sources to ensure pixel-level registration between data sources;

[0087] Perform atmospheric correction on optical remote sensing images to eliminate atmospheric scattering and absorption effects, and perform topographic correction on mountain grassland areas to remove topographic shadows and aspect effects;

[0088] Combine high-spatial-resolution and high-temporal-resolution data to construct a data cube with a unified time step and multiple spatial scales, and ensure spectral consistency among data from different sensors through radiometric calibration;

[0089] Based on a stratified sampling strategy, construct a ground quadrat network covering different grassland types and degradation degrees, collect data including vegetation cover, species composition, biomass, and soil physical and chemical properties, and form a standardized grassland survey dataset;

[0090] Conduct a joint analysis of ground measured data and high-resolution remote sensing images to establish an association database for grassland types and degradation degrees.

[0091] In summary, through refined data processing and the combination of multi-source remote sensing data, the accuracy and spatio-temporal adaptability of grassland monitoring have been significantly improved. First, grassland mapping is carried out by obtaining high-resolution optical satellite images, medium- and low-resolution optical satellite images are combined for long-term dynamic monitoring, and microwave remote sensing data is used to penetrate clouds to observe grassland biomass and moisture conditions, providing multi-dimensional data support for the comprehensive monitoring of grassland degradation. All images are processed by stripe removal, broken line repair, and denoising to ensure data quality. Second, geometric precision rectification of remote sensing images is carried out based on a high-precision control point network to ensure pixel-level registration between different data sources, thereby improving the spatial consistency of data from different sensors. In terms of image processing, the atmospheric effect is eliminated through atmospheric correction of optical remote sensing images, and terrain correction is carried out for mountain grassland areas to remove terrain shadows and aspect effects, thereby improving the extraction accuracy of ground information. In addition, by combining high-spatial-resolution and high-temporal-resolution data, a data cube with a unified time step and multiple spatial scales is constructed, and spectral consistency between different sensors is ensured through radiometric calibration, providing stable data support for long-term monitoring. Based on a stratified sampling strategy, a ground quadrat network covering different grassland types and degradation degrees is constructed, and data including vegetation cover, species composition, biomass, and soil physical and chemical properties are collected to form a standardized grassland survey dataset. Finally, through the joint analysis of ground measured data and high-resolution remote sensing images, an association database for grassland types and degradation degrees is established, providing high-precision and highly consistent basic data for grassland degradation monitoring and management.

[0092] Furthermore, the atmospheric correction of optical remote sensing images uses an improved 6S or FLAASH atmospheric radiative transfer model, and the parameter settings include: atmospheric visibility of 20 - 50 km, relative humidity of 40 - 70%, aerosol optical thickness of 0.05 - 0.5, and ozone content of 250 - 350 Dobson units; the correction process combines near-surface meteorological station data to ensure that the correction accuracy RMSE is less than 0.05 reflectance units.

[0093] Further, the ground quadrat network based on the stratified sampling strategy is designed as follows: 8-15 main quadrats are set for each grassland type, and 3-5 secondary quadrats are set within each main quadrat; the main quadrat is 30 m × 30 m in size, matching the pixel size of the medium-resolution remote sensing data; the secondary quadrat is 1 m × 1 m in size, and each secondary quadrat is arranged at an interval of 5-10 m; the survey indicators of the quadrats include vegetation coverage, dominant species composition, aboveground biomass, species richness, and soil surface structure.

[0094] Further, step 2 includes the following steps:

[0095] Based on the corrected multi-source remote sensing data, calculate the vegetation index, water index, and soil index, and construct a comprehensive index set that can comprehensively characterize the structural and functional characteristics of the grassland ecosystem;

[0096] For each indicator, calculate its statistical characteristics in the time dimension and space dimension, construct a multi-dimensional feature space, and accurately depict the dynamic changes of the grassland ecosystem at different degradation stages;

[0097] At the three spatial scales of pixel, patch, and landscape, calculate the information gain value and mutual information value of each indicator for grassland degradation classification, quantitatively evaluate the classification contribution degree of each indicator at different scales, and screen out a candidate indicator set with significant and stable contributions through information stability analysis;

[0098] Evaluate the response sensitivity of each indicator to the grassland degradation gradient, calculate the sensitivity coefficient and discrimination index, and combined with the true value library data, analyze the performance differences of the indicators in different ecological regions, grassland types, and degradation degree intervals, and identify the most discriminative characteristic indicators at each scale;

[0099] Automatically screen the optimal indicator combination at different spatial scales, ensure the classification accuracy while minimizing the redundancy between indicators, and comprehensively consider the economy and practicability of data acquisition to form an adaptive indicator combination scheme suitable for different scales and grassland types;

[0100] Establish cross-scale indicator transfer relationships and conversion rules, evaluate the classification performance and stability of the indicator systems at each scale based on an independent validation data set, and finally construct a grassland monitoring indicator system with scale self-adaptability, regional pertinence, and degradation sensitivity.

[0101] In summary, through precise multi-source remote sensing data analysis and feature extraction, the accuracy and adaptability of grassland degradation monitoring have been significantly improved. First, based on the corrected remote sensing data, vegetation indices, water indices, and soil indices were calculated, and a comprehensive set of comprehensive indicators representing the structure and function of the grassland ecosystem was constructed, providing multi-dimensional basic data for the monitoring of grassland ecological changes. Subsequently, for each indicator, statistical features in the time and space dimensions were calculated, and a multi-dimensional feature space was constructed to accurately describe the dynamic changes of the grassland ecosystem at different degradation stages. In terms of spatial scale analysis, the method quantitatively evaluated the classification contribution of each indicator at different scales by calculating the information gain value and mutual information value of each indicator for grassland degradation classification, and a candidate indicator set with significant and stable contributions was selected through information stability analysis. In addition, the technology also evaluated the response sensitivity of each indicator to the grassland degradation gradient, calculated the sensitivity coefficient and discrimination index, and analyzed the performance differences of the indicators in different ecological regions, grassland types, and degradation degree intervals in combination with the true value database data, so as to identify the most discriminative feature indicators. Finally, the method automatically selects the optimal indicator combination at different spatial scales, ensures the classification accuracy, minimizes redundant information, and comprehensively considers the economy and practicality of data acquisition to form an adaptive indicator combination scheme. By establishing cross-scale indicator transfer relationships and conversion rules, and evaluating the classification performance and stability in combination with an independent verification data set, a grassland monitoring indicator system with scale self-adaptability, regional pertinence, and degradation sensitivity was finally constructed.

[0102] Furthermore, for each indicator, calculate its statistical features in the time dimension and space dimension, construct a multi-dimensional feature space, and accurately depict the dynamic changes of the grassland ecosystem at different degradation stages, including the following steps:

[0103] For each indicator, calculate the time mean μ t , the time variance σ t 2 , the change rate V t and the autocorrelation R t to capture the stability, volatility, and trend information of the grassland ecosystem at different times;

[0104] At each moment, calculate the spatial mean μ s , the spatial variance σ s 2 and the spatial autocorrelation I s for all indicator data to quantify the balance, discreteness, and adjacent relationship of the ecosystem in spatial distribution;

[0105] Combine the statistical features extracted from the time and space dimensions into a multi-dimensional feature vector F = [μ t , σ t2 , V t , R t , μ s , σ s 2 , I s , a feature space for comprehensively depicting the dynamic changes of grassland degradation is formed;

[0106] By comparing the eigenvectors of different periods or regions and calculating the ecosystem dynamic change index D, the accurate determination and dynamic monitoring of the degradation stage of the grassland ecosystem are realized. The formula is expressed as: Among them, μ t (t) is the mean value of the index in the time dimension, representing the mean value of the grassland degradation index calculated at time point t; μ s (s) is the mean value of the index in the spatial dimension, the mean value of the grassland degradation index calculated in spatial unit s; μ t (t + 1) is the mean value of the index in the time dimension, the mean value of the grassland degradation index calculated at time point t + 1; μ s (s + 1) is the mean value of the index in the spatial dimension, the mean value of the grassland degradation index calculated in spatial unit s + 1.

[0107] Furthermore, at the three spatial scales of pixel, patch and landscape, calculate the information gain value and mutual information value of each index for grassland degradation classification, quantitatively evaluate the classification contribution degree of each index at different scales, and screen out the candidate index set with significant and stable contributions through information stability analysis, including the following steps:

[0108] According to the remote sensing data of the target area, divide the three spatial scales of pixel, patch and landscape;

[0109] Under each spatial scale, collect the corresponding grassland degradation classification labels and various remote sensing index data to ensure that the data set under each scale contains sufficient sample size and spatial information;

[0110] For each spatial scale, use the information gain formula IG(T, A) = H(T) - H(T|A) to calculate the information gain value of each remote sensing index A for grassland degradation classification T. Among them, IG(T, A) represents the information gain of remote sensing index A for grassland degradation classification T, that is, the degree of reduction in the uncertainty of grassland degradation classification under the condition of knowing A; H(T) represents the entropy of grassland degradation classification T, measuring the overall uncertainty of T; H(T|A) represents the conditional entropy of grassland degradation classification T after giving remote sensing index A, that is, the remaining uncertainty of T under the condition of knowing A;

[0111] Using the mutual information formula MI(A,T)=H(T)+H(A)-H(T,A), calculate the mutual information value between the remote sensing indicators and the grassland degradation classification at different spatial scales. Here, MI(A,T) represents the mutual information between the remote sensing indicator A and the grassland degradation classification T; H(A) represents the entropy of the remote sensing indicator A; H(T,A) represents the joint entropy of the grassland degradation classification T and the remote sensing indicator A.

[0112] Quantitatively evaluate the classification contribution degree of each remote sensing indicator at different spatial scales through the information gain value and the mutual information value.

[0113] Calculate the standard deviation of the information gain or mutual information value of the remote sensing indicator at different spatial scales, and use the formula: Evaluate the stability of the information, where μ IG represents the mean value of the information gain value calculated at different spatial scales, and is used to evaluate the overall classification contribution degree of the indicator; σ IG represents the standard deviation of the information gain value calculated at different spatial scales, and is used to measure the stability of the classification contribution degree of the indicator at each scale.

[0114] According to the results of the information stability analysis, screen out the remote sensing indicator set that is stable and has significant contribution at different scales.

[0115] Furthermore, step 3 includes the following steps:

[0116] Integrate the climate factor and human factor data to establish a spatio-temporally consistent potential driving force database for grassland degradation, and ensure that the data of each driving factor matches the remote sensing monitoring data in terms of time and space resolution.

[0117] Taking the grassland degradation index as the response variable and the climate and human factors as the predictive variables, construct a multi-level Bayesian network model, where: the first level represents the pixel-scale response, the second level represents the patch-scale response, and the third level represents the landscape-scale response. Each level is connected through conditional probability relationships.

[0118] Calculate the main effects, interaction effects of the climate and human factors and their contribution rates at different scales, and use partial correlation analysis and variance decomposition techniques to strip the independent contributions of each factor.

[0119] Combined with the information criterion and Bayesian factor comparison, identify the spatial scale conversion threshold of the dominant effects of the climate and human factors, and use the entropy maximization criterion to determine the optimal scale segmentation point, and construct a scale-driving force response curve to identify the key scale nodes.

[0120] Calculate the relative contribution rate of each driving factor to grassland degradation and its spatial differentiation characteristics at different scales, generate a scale-driving force contribution rate map, and realize the accurate quantification and spatial visualization of the impacts of the climate and human factors.

[0121] In summary, by integrating climate and anthropogenic factor data and combining remote sensing monitoring data, a spatially and temporally consistent potential driving force database for grassland degradation has been successfully constructed, thus ensuring the matching of the data of each driving factor with the remote sensing data in terms of temporal and spatial resolution. By constructing a multi-level Bayesian network model, this method accurately simulates the responses of grassland degradation at different spatial scales (pixel, patch, landscape), where each level of the model is connected through conditional probability relationships. Further, the method calculates the main effects, interaction effects of climate and anthropogenic factors and their contribution rates at different scales, and uses partial correlation analysis and variance decomposition techniques to strip the independent contributions of each factor, improving the accuracy of the analysis. By comparing information criteria and Bayesian factors, the spatial scale conversion thresholds of the dominant effects of climate and anthropogenic factors are identified, and the optimal scale segmentation points are determined using the entropy maximization criterion, constructing a scale-driving force response curve and identifying key scale nodes. In addition, the method also calculates the relative contribution rates of each driving factor to grassland degradation and their spatial differentiation characteristics at different scales, generates a scale-driving force contribution rate map, and realizes the precise quantification and spatial visualization of the impacts of climate and anthropogenic factors.

[0122] Furthermore, climate factors include average temperature, precipitation, evapotranspiration and extreme climate events; anthropogenic factors include grazing pressure, agricultural and pastoral activities, infrastructure construction and protection policies.

[0123] Furthermore, the construction process of the multi-level Bayesian network model includes: in the structure learning stage, the network topology structure is determined by combining the greedy search algorithm and expert knowledge constraints; in the parameter learning stage, the expectation maximization algorithm is used to estimate the parameters of the conditional probability table; the evaluation indicators of the network model include the AUC value, Kappa coefficient and average log-likelihood value; the training-validation partition ratio is 7:3, and k-fold cross-validation is used to evaluate the model stability.

[0124] Furthermore, the identification of the key scale nodes adopts the following criteria: calculate the Bayesian factor B at each candidate scale, and the formula is: B = P(D|M1) / P(D|M2), where M1 is the climate factor-dominated model; M2 is the anthropogenic factor-dominated model; P(D|M1) represents the probability of the occurrence of the observed data D under the climate factor-dominated model M1; P(D|M2) represents the probability of the occurrence of the observed data D under the anthropogenic factor-dominated model M2; when 2 ≤ |logB| ≤ 6, it is medium evidence, and when |logB| > 6, it is strong evidence, and the scale point with the largest |logB| value is used as the key conversion node; at the same time, multi-scale variance analysis is combined to verify the significance of the scale conversion point.

[0125] Furthermore, calculate the main effects, interaction effects of climate and anthropogenic factors and their contribution rates at different scales, and use partial correlation analysis and variance decomposition techniques to separate the independent contributions of each factor, including the following steps:

[0126] Use a multiple linear regression model, set the degree of grassland degradation as the dependent variable, and climate factors and anthropogenic factors as independent variables. Obtain the main effect coefficients through regression analysis and calculate the contribution rate of each main effect to the total variance;

[0127] Add the interaction term of climate factors and anthropogenic factors to the regression model, calculate the interaction effect coefficient, and calculate the contribution rate of the interaction effect through analysis of variance to measure the impact of the interaction between climate and anthropogenic factors on grassland degradation;

[0128] By calculating the partial correlation coefficient and respectively evaluate the independent contributions of climate and anthropogenic factors to grassland degradation under the condition of controlling the influence of another factor. Among them, represents the partial correlation coefficient between variable X1 and dependent variable Y after controlling variable X2; represents the partial correlation coefficient between variable X2 and dependent variable Y after controlling variable X1;

[0129] Apply the variance decomposition method to decompose the total variance into the variances of climate factors, anthropogenic factors, interaction effects and error terms;

[0130] By calculating the variance contribution rate of each part, clarify the contribution degree of each factor to the variation of grassland degradation.

[0131] Furthermore, step 4 includes the following steps:

[0132] Adopt an improved convolutional neural network to extract local texture features at the pixel scale, integrate an attention mechanism at the patch scale to enhance the feature expression of key regions, and introduce a graph convolutional network at the landscape scale to capture the topological relationship between landscape units;

[0133] Utilize spatial context information to optimize the classification boundary and achieve high-precision grassland degradation level recognition at a resolution of 1–30 meters;

[0134] Perform multi-scale segmentation on high-resolution remote sensing images to generate homogeneous patches with ecological significance, realize the degradation classification at the grassland functional unit level, and analyze the impact of grassland fragmentation and connectivity changes on ecological functions;

[0135] Capture the macroscopic pattern and functional changes of the grassland ecosystem, analyze the spatial aggregation and diffusion trends of degraded areas, and evaluate the integrity of the ecosystem;

[0136] Based on the multi - period classification results, a degradation transition matrix is constructed to quantitatively calculate the area proportion, spatial distribution characteristics, and their temporal evolution laws of each degradation level at different scales.

[0137] In summary, by introducing advanced deep learning and graph neural network technologies, the accuracy and spatio - temporal adaptability of grassland degradation level recognition have been significantly improved. At the pixel scale, an improved convolutional neural network is used to extract local texture features. At the patch scale, an attention mechanism is integrated to enhance the expression of key region features. At the landscape scale, by introducing a graph convolutional network to capture the topological relationship between landscape units, the spatial structure of the grassland ecosystem can be better characterized. Using spatial context information to optimize the classification boundary, this method achieves high - precision grassland degradation level recognition at a resolution of 1 - 30 meters. At the same time, by multi - scale segmentation of high - resolution remote sensing images, homogeneous patches with ecological significance are generated, further realizing the degradation classification at the grassland functional unit level, and analyzing the impact of grassland fragmentation and connectivity changes on ecological functions. This method can capture the macroscopic pattern and functional changes of the grassland ecosystem, analyze the spatial aggregation and diffusion trends of degraded areas, and thus evaluate the integrity of the ecosystem. In addition, based on the multi - period classification results, a degradation transition matrix is constructed to quantitatively calculate the area proportion, spatial distribution characteristics, and their temporal evolution laws of each degradation level at different scales, providing comprehensive quantitative analysis support for grassland protection and restoration.

[0138] Furthermore, the improved convolutional neural network is based on the ResNet50 or DenseNet121 architecture, and combines depth - wise separable convolutions to reduce the number of network parameters by 40 - 60%; the network input includes remote sensing images of 10 - 12 bands and 3 - 5 terrain - derived indicators; during the training process, the Adam optimizer with a learning rate of 0.0005 - 0.002 is used, the batch size is 16 - 64, and the number of training epochs is 50 - 200. The cosine annealing learning rate strategy and early stopping method are used to avoid overfitting;

[0139] The attention mechanism uses a dual - channel and spatial attention module. Among them, the channel attention extracts the importance of channel features through the combination of global average pooling and max - pooling, and the spatial attention generates a spatial weight map through a 7×7 convolution; the two attention modules are applied in series to the feature map to improve the model's adaptive learning ability for different bands and spatial regions.

[0140] Furthermore, step 5 includes the following steps:

[0141] Establish a conversion rule library among the three scales of pixels, patches, and landscapes. Specifically: for discrete classification results, the dominant type method and weighted voting method are used for upscaling aggregation; for continuous degradation indices, the regional average method and area weighting method are used for scale conversion. At the same time, a non-linear conversion function considering spatial heterogeneity is designed to capture the abrupt change effect of degradation characteristics during scale conversion;

[0142] Use the global information at the high-level scale as prior knowledge to constrain the local detail expression at the lower-level scale. At the same time, use the high-precision information at the lower-level scale as evidence to update the overall cognition at the upper-level scale, and construct a two-way information flow mechanism;

[0143] Calculate the entropy value, variance, and confidence interval of each spatial unit at different scales, identify high-uncertainty regions, and analyze the main factors causing uncertainty;

[0144] Identify the grassland types and degradation processes that are highly sensitive to a specific scale, calculate the scale adaptability index, quantitatively evaluate the applicability of the monitoring results at each scale, and construct an optimal scale selection decision tree;

[0145] Fuse the monitoring results at each scale according to the reliability weights to generate a comprehensive grassland degradation monitoring result with a unified accuracy standard and complete spatial coverage.

[0146] In summary, by establishing a multi-scale conversion rule library, accurate conversion among the pixel, patch, and landscape scales is achieved, significantly improving the spatial adaptability and accuracy of grassland degradation monitoring. For discrete classification results, the dominant type method and weighted voting method are used for upscaling; for continuous degradation indices, the regional average method and area weighting method are used for scale conversion. At the same time, a non-linear conversion function considering spatial heterogeneity is designed to capture the abrupt change effect of degradation characteristics during scale conversion. This method also constructs a two-way information flow mechanism, using the global information at the high-level scale to constrain the local details at the lower-level scale, and using the high-precision information at the lower-level scale to update the overall cognition at the upper-level scale, ensuring the effective fusion of information at each scale. By calculating the entropy value, variance, and confidence interval of each spatial unit at different scales, identifying high-uncertainty regions, and analyzing the main factors causing uncertainty, the reliability of the monitoring results is further improved. This technology can identify the grassland types and degradation processes that are highly sensitive to a specific scale, calculate the scale adaptability index, quantitatively evaluate the applicability of the monitoring results at each scale, and finally construct an optimal scale selection decision tree. By fusing the monitoring results at each scale and weighting them according to the reliability weights, a comprehensive grassland degradation monitoring result with a unified accuracy standard and complete spatial coverage is generated.

[0147] Furthermore, the specific implementation of constructing the two-way information flow mechanism includes: the top-down constraint flow adopts a hierarchical conditional random field model for spatial context constraint; the bottom-up update flow adopts an aggregation-decomposition strategy to upload information through majority voting, regional statistics, and boundary-preserving filtering; the information flows in both directions are iterated alternately 2-3 times until the change rate of the classification results between adjacent iterations is less than 5%, or the maximum number of iterations, which is 5 times, is reached; the consistency index Kappa between scales is required to reach above 0.65.

[0148] Furthermore, step 6 includes the following steps:

[0149] Based on the grassland degradation level classification results and spatial statistical analysis, a high-precision degradation level distribution map is generated, and the spatial distribution characteristics of areas with different degradation degrees are visually displayed through a graded color scale.

[0150] Construct a spatial contribution rate model of climate factors and human factors, generate a driving force analysis map, and use a two-factor color coding method to visually present the spatial heterogeneity of the dominant factors.

[0151] Combining historical data sequences with current monitoring results, a grassland degradation risk warning map is generated to identify potential rapid degradation areas and ecological fragile points, and corresponding warning thresholds are set according to the risk levels.

[0152] According to the grassland degradation degree, driving factors, and ecological sensitivity, management units are divided, a zoning management recommendation map is generated, and targeted management measures are proposed to achieve the efficiency and pertinence of grassland protection and restoration.

[0153] In summary, through the grassland degradation level classification results and spatial statistical analysis, a high-precision degradation level distribution map is generated, visually displaying the spatial distribution characteristics of areas with different degradation degrees. Combining the spatial contribution rate model of climate factors and human factors, a driving force analysis map is generated, and the spatial heterogeneity of the dominant factors is clearly presented through a two-factor color coding method. By combining historical data and current monitoring results, a grassland degradation risk warning map is generated, which can identify potential rapid degradation areas and ecological fragile points and provide warning thresholds for risk management. In addition, based on the grassland degradation degree, driving factors, and ecological sensitivity, management units are divided and a zoning management recommendation map is generated, and targeted management measures are proposed, providing strong support for achieving the efficiency and pertinence of grassland protection and restoration.

[0154] As Figure 2 shown, the grassland degradation degree monitoring model based on remote sensing technology includes the following modules:

[0155] The data acquisition and preprocessing module is responsible for integrating multi-source remote sensing data and ground measured data, constructing a spatio-temporally consistent multi-scale dataset, and performing preprocessing to ensure the accuracy and consistency of the data;

[0156] The degradation degree assessment module, based on the preprocessed remote sensing data, calculates relevant indicators and conducts multi-dimensional feature analysis to screen out the optimal indicator combination;

[0157] The driving force analysis module quantifies the main effects, interaction effects and their contribution rates of climate and human factors on grassland degradation, and identifies key scale nodes;

[0158] The result display and management module displays the grassland degradation level distribution, driving force analysis, risk warning and management suggestions in an intuitive chart form and conducts management;

[0159] The multi-scale analysis and conversion module establishes conversion rules between different spatial scales, realizes multi-scale segmentation and feature extraction, constructs a two-way information flow mechanism, and ensures the reliable conversion and fusion of analysis results at different scales;

[0160] The management strategy generation module automatically generates differentiated management strategies according to the grassland degradation monitoring results and relevant analyses, proposes targeted management measures, and realizes the efficiency and pertinence of grassland protection and restoration.

[0161] In summary, the grassland degradation degree monitoring model based on remote sensing technology realizes the precise monitoring and management of grassland degradation through the collaborative work of multiple modules. The data acquisition and preprocessing module integrates multi-source remote sensing and ground data, constructs a spatio-temporally consistent multi-scale dataset, and ensures data accuracy and consistency. The degradation degree assessment module screens out the optimal remote sensing indicators through multi-dimensional feature analysis and accurately evaluates the grassland degradation degree. The driving force analysis module quantifies the main and interaction effects of climate and human factors on degradation and identifies key scale nodes. The result display and management module presents the grassland degradation level, driving force analysis, risk warning and management suggestions in an intuitive chart, providing a basis for decision-making. The multi-scale analysis and conversion module realizes the conversion and fusion of analysis results at different spatial scales and constructs a two-way information flow mechanism to ensure the effective transmission of information between scales. The management strategy generation module automatically generates differentiated management strategies according to the monitoring results and proposes efficient and targeted grassland protection and restoration measures.

[0162] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than limiting it. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for monitoring the degree of grassland degradation based on remote sensing technology, characterized in that, It includes the following steps: Step 1: Construct a spatio-temporally consistent multi-scale remote sensing dataset, and combine it with ground measured data to establish a high-precision true value library for grassland types and degradation degrees; Step 2: Automatically screen the remote sensing indicators most sensitive to grassland degradation, and construct a remote sensing indicator system adapted to different spatial scales; Step 3: Identify the key scale conversion nodes where grassland degradation is driven by climate and human factors; Step 4: Realize the classification of grassland degradation degrees from pixel scale to landscape scale, and quantitatively analyze the spatial distribution, area change and spatio-temporal evolution trend of each degradation level; Step 5: Realize the reliable conversion and up-down transfer of analysis results at different scales, and generate comprehensive grassland degradation monitoring results; Step 6: Automatically generate differentiated management strategies, and output the grassland degradation level distribution map, driving force analysis map, risk warning map and zoning management suggestion map.

2. The method for monitoring the degree of grassland degradation based on remote sensing technology according to claim 1, wherein Step 2 includes the following steps: Based on the corrected multi-source remote sensing data, calculate the vegetation index, water index and soil index, and construct a comprehensive index set that can comprehensively characterize the structural and functional characteristics of the grassland ecosystem; For each indicator, calculate its statistical characteristics in the time dimension and space dimension, construct a multi-dimensional feature space, and accurately depict the dynamic changes of the grassland ecosystem at different degradation stages; At the three spatial scales of pixel, patch and landscape, calculate the information gain value and mutual information value of each indicator for grassland degradation classification, quantitatively evaluate the classification contribution degree of each indicator at different scales, and screen out the candidate indicator set with significant and stable contributions through information stability analysis; Evaluate the response sensitivity of each indicator to the grassland degradation gradient, calculate the sensitivity coefficient and discrimination index, and combined with the true value library data, analyze the performance differences of the indicators in different ecological regions, grassland types and degradation degree intervals, and identify the most discriminative characteristic indicators at each scale; Automatically screen the optimal indicator combination at different spatial scales, minimize the redundancy between indicators while ensuring the classification accuracy, and comprehensively consider the economy and practicability of data acquisition to form an adaptive indicator combination scheme adapted to different scales and grassland types; Establish cross-scale indicator transfer relationships and conversion rules, evaluate the classification performance and stability of each scale indicator system based on an independent verification dataset, and finally construct a grassland monitoring indicator system with scale self-adaptability, regional pertinence and degradation sensitivity.

3. The grassland degradation degree monitoring method based on remote sensing technology according to claim 2, characterized in that, For each indicator, calculate its statistical characteristics in the time dimension and space dimension, construct a multi-dimensional feature space, and accurately depict the dynamic changes of the grassland ecosystem at different degradation stages, including the following steps: Calculate the time mean μ for each indicator t , the time variance σ t 2 , the change rate V t and the autocorrelation R t , to capture the stability, volatility and trend information of the grassland ecosystem at different times; At each moment, calculate the spatial mean μ of all the index data s , the spatial variance σ s 2 and the spatial autocorrelation I s , to quantify the equilibrium, discreteness and adjacent relationship of the ecosystem in spatial distribution; Combine the statistical features extracted from the time and space dimensions into a multi-dimensional feature vector F = [μ t , σ t 2 , V t , R t , μ s , σ s 2 , I s , forming a feature space that comprehensively characterizes the dynamic changes of grassland degradation; By comparing the eigenvectors at different times or regions and calculating the ecosystem dynamic change index D, the accurate determination and dynamic monitoring of the degradation stage of the grassland ecosystem are realized. The formula is expressed as: Among them, μ t (t) is the mean value of the index in the time dimension, representing the mean value of the grassland degradation index calculated at time point t; μ s (s) is the mean value of the index in the spatial dimension, the mean value of the grassland degradation index calculated in the spatial unit s; μ t (t + 1) is the mean value of the index in the time dimension, the mean value of the grassland degradation index calculated at time point t + 1; μ s (s + 1) is the mean value of the index in the spatial dimension, the mean value of the grassland degradation index calculated in the spatial unit s + 1.

4. The method for monitoring the degree of grassland degradation based on remote sensing technology according to claim 2, characterized in that, At the three spatial scales of pixel, patch and landscape, calculate the information gain value and mutual information value of each indicator for grassland degradation classification, quantitatively evaluate the classification contribution degree of each indicator at different scales, and screen out the candidate indicator set with significant and stable contributions through information stability analysis, including the following steps: According to the remote sensing data of the target area, divide the three spatial scales of pixel, patch and landscape; Under each spatial scale, collect the corresponding grassland degradation classification labels and various remote sensing indicator data to ensure that the dataset under each scale contains sufficient sample size and spatial information; For each spatial scale, the information gain formula IG(T,A) = H(T) - H(T|A) is used to calculate the information gain value of each remote sensing index A for the grassland degradation classification T. Here, IG(T,A) represents the information gain of the remote sensing index A for the grassland degradation classification T, that is, the degree of reduction in the uncertainty of the grassland degradation classification under the condition of knowing A; H(T) represents the entropy of the grassland degradation classification T, measuring the overall uncertainty of T; H(T|A) represents the conditional entropy of the grassland degradation classification T after the given remote sensing index A, that is, the remaining uncertainty of T under the condition of knowing A. The mutual information formula MI(A,T) = H(T) + H(A) - H(T,A) is used to calculate the mutual information value between the remote sensing index and the grassland degradation classification at different spatial scales. Here, MI(A,T) represents the mutual information between the remote sensing index A and the grassland degradation classification T; H(A) represents the entropy of the remote sensing index A; H(T,A) represents the joint entropy of the grassland degradation classification T and the remote sensing index A. Through the information gain value and the mutual information value, a quantitative evaluation of the classification contribution of each remote sensing index at different spatial scales is carried out. Calculate the standard deviation of the information gain or mutual information value of remote sensing indicators at different spatial scales, and use the formula: Evaluate the stability of the information, where μ IG represents the mean of the information gain values calculated at different spatial scales and is used to evaluate the overall classification contribution of the indicator; σ IG represents the standard deviation of the information gain values calculated at different spatial scales and is used to measure the stability of the classification contribution of the indicator at each scale; According to the results of the information stability analysis, a set of remote sensing indices that are stable and have significant contributions at different scales is selected.

5. The method for monitoring the degree of grassland degradation based on remote sensing technology according to claim 1, wherein, Step 3 includes the following steps: Integrate climate factor and human factor data to establish a spatio-temporally consistent potential driving force database for grassland degradation, ensuring that the data of each driving factor matches the remote sensing monitoring data in terms of time and space resolution. Taking the grassland degradation index as the response variable and climate and human factors as the predictive variables, construct a multi-level Bayesian network model, where: the first level represents the pixel-scale response, the second level represents the patch-scale response, and the third level represents the landscape-scale response, and the levels are connected through conditional probability relationships. Calculate the main effects, interaction effects of climate and human factors and their contribution rates at different scales, and use partial correlation analysis and variance decomposition techniques to separate the independent contributions of each factor. Combined with the information criterion and Bayesian factor comparison, identify the spatial scale conversion threshold of the dominant effects of climate and human factors, and use the entropy maximization criterion to determine the optimal scale segmentation point, construct a scale-driving force response curve to identify key scale nodes. Calculate the relative contribution rates of each driving factor to grassland degradation and their spatial differentiation characteristics at different scales, generate a scale-driving force contribution rate map, and realize the accurate quantification and spatial visualization of the impacts of climate and human factors.

6. The method for monitoring the degree of grassland degradation based on remote sensing technology according to claim 5, wherein Calculate the main effects, interaction effects of climate and human factors and their contribution rates at different scales, and use partial correlation analysis and variance decomposition techniques to separate the independent contributions of each factor, including the following steps: Use a multiple linear regression model, set the grassland degradation degree as the dependent variable, and climate factors and human factors as the independent variables, and obtain the main effect coefficients through regression analysis and calculate the contribution rate of each main effect to the total variance. Add the interaction term of climate factors and human factors to the regression model, calculate the interaction effect coefficient, and calculate the contribution rate of the interaction effect through variance analysis to measure the impact of the interaction between climate and human factors on grassland degradation. By calculating the partial correlation coefficient and respectively evaluate the independent contributions of climate and anthropogenic factors to grassland degradation under the condition of controlling the influence of another factor. Among them, represents the partial correlation coefficient between variable X1 and dependent variable Y after controlling variable X2; represents the partial correlation coefficient between variable X2 and dependent variable Y after controlling variable X1; Using the variance decomposition method, decompose the total variance into the variances of climate factors, human factors, interaction effects, and error terms; By calculating the variance contribution rate of each part, clarify the contribution degree of each factor to the variation of grassland degradation.

7. The method for monitoring the degree of grassland degradation based on remote sensing technology according to claim 1, characterized in that Step 4 includes the following steps: At the pixel scale, adopt an improved convolutional neural network to extract local texture features. At the patch scale, integrate an attention mechanism to enhance the feature expression of key areas. At the landscape scale, introduce a graph convolutional network to capture the topological relationships between landscape units; Utilize spatial context information to optimize the classification boundary and achieve high-precision grassland degradation level recognition at a resolution of 1–30 meters; Perform multi-scale segmentation on high-resolution remote sensing images to generate homogeneous patches with ecological significance, achieve degradation grading at the grassland functional unit level, and analyze the impacts of grassland fragmentation and connectivity changes on ecological functions; Capture the macroscopic pattern and functional changes of the grassland ecosystem, analyze the spatial aggregation and diffusion trends of degraded areas, and evaluate the integrity of the ecosystem; Based on the classification results of multiple periods, construct a degradation transition matrix, and quantitatively calculate the area proportion, spatial distribution characteristics, and temporal evolution laws of each degradation level at different scales.

8. The method for monitoring the grassland degradation degree based on remote sensing technology according to claim 1, characterized in that, Step 5 includes the following steps: Establish a conversion rule library between the pixel, patch, and landscape scales. Among them: for discrete classification results, use the dominant type method and weighted voting method for upscaling aggregation; for continuous degradation indices, use the regional average method and area weighting method for scale conversion; at the same time, design a non-linear conversion function considering spatial heterogeneity to capture the mutation effect of degradation characteristics during scale conversion; Use the global information at the high-level scale as prior knowledge to constrain the local detail expression at the lower-level scale. At the same time, use the high-precision information at the lower-level scale as evidence to update the overall cognition at the upper-level scale, and construct a two-way information flow mechanism; Calculate the entropy value, variance, and confidence interval of each spatial unit at different scales, identify high-uncertainty regions, and analyze the main factors causing uncertainty; Identify the grassland types and degradation processes that are highly sensitive to a specific scale, calculate the scale adaptability index, quantitatively evaluate the applicability of the monitoring results at each scale, and construct an optimal scale selection decision tree; Fuse the monitoring results at each scale according to the reliability weights to generate a comprehensive grassland degradation monitoring result with a unified accuracy standard and complete spatial coverage.

9. The method for monitoring the degree of grassland degradation based on remote sensing technology according to claim 1, wherein, Step 6 includes the following steps: Based on the grassland degradation level classification results and spatial statistical analysis, generate a high-precision degradation level distribution map, and visually display the spatial distribution characteristics of areas with different degradation degrees through a graded color scale; Construct a spatial contribution rate model of climate factors and human factors, generate a driving force analysis map, and use a two-factor color coding method to visually present the spatial heterogeneity of the dominant factors; Combine historical data sequences with current monitoring results to generate a grassland degradation risk warning map, identify potential rapidly degraded areas and ecological vulnerable points, and set corresponding warning thresholds according to the risk level; According to the grassland degradation degree, driving factors, and ecological sensitivity, divide management units, generate a zoning management recommendation map, and propose targeted management measures to achieve the efficiency and pertinence of grassland protection and restoration.

10. A monitoring model for the degree of grassland degradation based on remote sensing technology, which is used to implement the monitoring method for the degree of grassland degradation based on remote sensing technology according to any one of claims 1-9, characterized in that, It includes the following modules: The data acquisition and preprocessing module is responsible for integrating multi-source remote sensing data and ground measured data, constructing a spatio-temporally consistent multi-scale dataset, and performing preprocessing to ensure the accuracy and consistency of the data; The degradation degree assessment module calculates relevant indicators and conducts multi-dimensional feature analysis based on the preprocessed remote sensing data to screen out the optimal indicator combination; The driving force analysis module quantifies the main effects, interaction effects and their contribution rates of climate and human factors on grassland degradation, and identifies key scale nodes; The result display and management module displays the grassland degradation level distribution, driving force analysis, risk warning and management suggestions in an intuitive chart form and conducts management; The multi-scale analysis and conversion module establishes conversion rules between different spatial scales, realizes multi-scale segmentation and feature extraction, constructs a two-way information flow mechanism to ensure the reliable conversion and fusion of analysis results at different scales; The management strategy generation module automatically generates differentiated management strategies according to the grassland degradation monitoring results and relevant analyses, proposes targeted management measures, and realizes the efficiency and pertinence of grassland protection and restoration.

Images