Remote sensing monitoring method and system for grassland biomass

Abstract

The invention discloses a remote sensing monitoring method and system for grassland biomass, and the method comprises the following steps: constructing a high-dimensional feature matrix representing the state of a grassland ecosystem through employing a multi-mode remote sensing data fusion technology; based on the high-dimensional feature matrix and the dynamic spatio-temporal feature analysis framework, generating spatio-temporal correlation parameters of the grassland biomass; in combination with a deep neural network and a multi-task optimization algorithm, training a nonlinear biomass prediction model; carrying out conjoint analysis on a prediction model result and a high-resolution image, and carrying out hierarchical processing on a heterogeneous grassland region by applying a region segmentation technology; performing uncertainty propagation analysis on the partition estimation result and the multi-modal data, and outputting an optimized grassland biomass estimation result; through a cooperative monitoring mechanism of the unmanned aerial vehicle and satellite remote sensing, combining a partition estimation result with cross-spatio-temporal scale data, and dynamically updating spatio-temporal correlation parameters; and according to the optimized monitoring data, generating a grassland biomass spatial distribution map and a dynamic change trend analysis result.

Description

Technical Field

[0001] The present invention relates to the technical field of remote sensing monitoring, and in particular to a remote sensing monitoring method and system for grassland biomass. Background Art

[0002] With the rapid development of remote sensing technology and ecological monitoring needs, grassland biomass monitoring has become an important part of precision agriculture and ecological management. As an important indicator of grassland ecosystems, grassland biomass can not only reflect grassland productivity, but also provide information on grassland health and its dynamic changes. However, due to the spatial heterogeneity and spatiotemporal dynamics of grasslands, existing remote sensing monitoring technologies for grassland biomass face many challenges in terms of accuracy and applicability.

[0003] In the existing technology, traditional grassland biomass remote sensing monitoring methods mostly rely on a single data source or estimation methods based on simple empirical models. These methods often show significant limitations when faced with complex terrain, diverse grassland ecological environments, and dynamically changing biomass characteristics. Specifically, the existing methods have shortcomings in the following aspects:

[0004] 1. Insufficient data fusion capability: Existing technologies usually rely on a single remote sensing data source or simple multi-source data superposition, lacking an effective data fusion mechanism, resulting in a single dimension of monitoring data that cannot fully reflect the complex characteristics of grassland ecosystems;

[0005] 2. Low model accuracy: Traditional grassland biomass estimation models are mostly based on linear regression or simple exponential formulas, which fail to fully consider nonlinear factors and grassland heterogeneity, resulting in low model adaptability and prediction accuracy;

[0006] 3. Inadequate response to dynamic changes: Existing methods lack in-depth analysis of the spatiotemporal dynamic characteristics of grassland biomass and cannot accurately capture seasonal changes and heterogeneous distribution characteristics, resulting in insufficient reliability of monitoring results in dynamic assessment;

[0007] 4. Weak uncertainty processing capabilities: Traditional remote sensing monitoring methods lack effective analysis of the quantification and propagation of uncertainty in multi-source data, making it difficult to assess the reliability of monitoring results and optimize the data processing process;

[0008] 5. The contradiction between spatial resolution and coverage: Existing technologies make it difficult to balance the needs of high spatial resolution and large-scale coverage. Simply relying on drone or satellite data cannot achieve detailed monitoring and comprehensive coverage of grassland biomass.

[0009] Therefore, how to provide a remote sensing monitoring method and system for grassland biomass is an urgent problem to be solved by those skilled in the art. Summary of the invention

[0010] One purpose of the present invention is to propose a remote sensing monitoring method and system for grassland biomass. The present invention adopts technologies such as multimodal remote sensing data fusion, deep learning neural network modeling and dynamic spatiotemporal feature analysis, and describes in detail the specific steps for achieving accurate estimation and dynamic monitoring of grassland biomass. It has the advantages of high data fusion efficiency, high model prediction accuracy and strong dynamic change response capability.

[0011] A remote sensing monitoring method and system for grassland biomass according to an embodiment of the present invention comprises the following steps:

[0012] S1. Using multimodal remote sensing data fusion technology, we integrate UAV images, high-resolution satellite images, and time-series remote sensing data to generate a cross-resolution, multi-temporal joint feature space, extract grassland spectral features, texture features, and geometric structure features, and construct a high-dimensional feature matrix that characterizes the state of grassland ecosystems.

[0013] S2. Based on the high-dimensional feature matrix and dynamic spatiotemporal feature analysis framework, the temporal change decomposition and spatial heterogeneity clustering methods are used to generate spatiotemporal correlation parameters of grassland biomass and construct a description set of grassland dynamic change characteristics;

[0014] S3. Using the dynamic change feature description set, combined with deep neural network and multi-task optimization algorithm, the nonlinear biomass prediction model is trained to achieve high-precision grassland biomass estimation based on multi-source data characteristics and grassland heterogeneity distribution;

[0015] S4. Combine the prediction model results with high-resolution images, apply regional segmentation technology to perform stratified processing on heterogeneous grassland areas, and generate refined grassland biomass zoning estimation results to capture the complex local changes of grasslands;

[0016] S5. Conduct uncertainty propagation analysis on the partition estimation results and multimodal data, use uncertainty quantification optimization algorithm to evaluate the error sources of monitoring data, and output optimized grassland biomass estimation results;

[0017] S6. Through the coordinated monitoring mechanism of UAV and satellite remote sensing, the zoning estimation results are combined with cross-temporal and spatial scale data, and the spatiotemporal correlation parameters are dynamically updated to optimize the real-time nature of grassland biomass monitoring results in the temporal dimension and the integrity in the spatial dimension;

[0018] S7. Based on the optimized monitoring data, generate high-resolution grassland biomass spatial distribution maps and dynamic change trend analysis results to form a comprehensive assessment of grassland ecosystem status and grassland biomass changes.

[0019] Optionally, the S1 specifically includes:

[0020] S11. Obtain UAV remote sensing image data, collect grassland surface detail information through multispectral sensors, generate high-resolution images, and simultaneously obtain multi-temporal high-resolution satellite image data and other remote sensing data covering the study area;

[0021] S12. Data preprocessing of UAV and satellite images, including radiometric correction, geometric correction and noise removal, to generate standardized image sets and make multi-source data comparable in spectral, geometric and temporal dimensions;

[0022] S13. Apply data registration algorithms to spatially align remote sensing data of different resolutions and multiple temporal phases, and use SIFT algorithm to geometrically register images to generate cross-resolution and multi-temporal remote sensing data sets;

[0023] S14. Extract multimodal features of remote sensing data, including spectral features, texture features and geometric structure features. The spectral features are calculated using the Normalized Difference Vegetation Index (NDVI) and the Enhanced Vegetation Index (EVI):

[0024]

[0025] Among them, NIR represents the reflectivity of the near infrared band, R represents the reflectivity of the red light band, and B represents the reflectivity of the blue light band;

[0026] S15, extracting texture features of the image, using a gray level co-occurrence matrix to calculate contrast, homogeneity and energy indexes of the image, and generating a multi-dimensional texture feature matrix;

[0027] S16, extracting geometric structural features of remote sensing images, combining watershed segmentation algorithm and edge detection algorithm to identify grassland terrain boundaries and morphological structures, and generating a geometric feature description set related to spatial distribution;

[0028] S17. Construct a high-dimensional feature matrix to characterize the state of grassland ecosystem, and fuse the extracted spectral features, texture features and geometric structure features into a joint feature vector.

[0029] Optionally, the S2 specifically includes:

[0030] S21. Using the time series remote sensing data in the high-dimensional feature matrix, a time series decomposition model is established, and the seasonal and trend change characteristics are extracted by the weighted moving average method:

[0031]

[0032] Among them, T(t) is the trend component of the time series, X(ti) is the observation value at the previous i moments, W(i) is the weighting factor, and n is the moving window length;

[0033] S22. Combining the spatial heterogeneity clustering method, the spatial features in the high-dimensional feature matrix are clustered by a density-based spatial clustering algorithm to generate spatial partitions within the grassland area;

[0034] S23. Analyze the dynamic characteristics of each spatial partition and construct a grassland biomass dynamic change model with time and space dimensions as the core:

[0035] M(x,t)=F(S(x),T(t),P(x,t));

[0036] Among them, M(x,t) represents the dynamic characteristics at position x and time t, S(x) is the spatial characteristics, T(t) is the temporal characteristics, and P(x,t) is the external influencing factor;

[0037] S24. Through the adaptive time window method, the spatiotemporal correlation parameters within the partition are dynamically updated, and a grassland dynamic feature description set is generated according to the seasonal fluctuations and heterogeneous distribution changes of grassland biomass.

[0038] Optionally, the S3 specifically includes:

[0039] S31, dividing the dynamically changing feature description set into a training data set and a validation data set, and mapping the spatiotemporal correlation parameters and external environmental variables to a unified numerical range through normalization processing to adapt to the input requirements of the deep neural network;

[0040] S32. Construct a multi-layer deep neural network structure, including an input layer, a hidden layer and an output layer, wherein the input layer receives the key features in the dynamic change feature description set, the hidden layer realizes nonlinear feature conversion through an activation function, and the output layer generates a grassland biomass prediction value:

[0041] y=f(W3·g(W2·g(W1·x+b1)+b2)+b3);

[0042] Where x is the input feature vector, W1, W2, W3 are weight matrices, b1, b2, b3 are bias vectors, g(·) is the activation function, and f(·) is the activation function of the output layer;

[0043] S33. Combined with the multi-task optimization algorithm, the loss function of the grassland biomass prediction task is defined. The loss function consists of the mean square error and the uncertainty quantization error:

[0044] L = α·MSE + β·UQE;

[0045] Among them, α and β are task weight parameters, UQE is the uncertainty measure of the predicted value;

[0046] S34, using the back propagation algorithm to optimize the network weights and bias parameters, adjusting the network structure according to the gradient update principle of the loss function, and gradually minimizing the prediction error;

[0047] S35, verifying the optimized model, and using the validation data set to calculate the accuracy indicators of the prediction results, including correlation coefficients and relative errors, to verify the applicability of the model under grassland heterogeneity distribution and multi-source data characteristics;

[0048] S36. The trained and verified deep neural network model is used as the final nonlinear biomass prediction model to generate the spatial distribution estimation results of grassland biomass.

[0049] Optionally, the S4 specifically includes:

[0050] S41. Perform spatial overlap analysis on the grassland biomass distribution results generated by the prediction model and the high-resolution remote sensing images to calibrate the spatial distribution consistency between the prediction results and the images;

[0051] S42. Based on the calibrated data, the heterogeneous grassland area is segmented using the regional segmentation technology, and the watershed algorithm is used to generate the initial segmentation area according to the spectral difference of the image and the terrain change characteristics;

[0052] S43, optimizing the initial segmented area, combining edge detection method and morphological operation to eliminate over-segmented and under-segmented areas, and the optimized segmented area is expressed in vector form as:

[0053] Ri={x|x∈Ω,g(x)>τi};

[0054] Among them, R i is the i-th segmentation region, x is a pixel point, Ω is the global research area, g(x) is the pixel feature function in the region, τ i is the segmentation threshold;

[0055] S44. Combine the segmented regions, analyze the biomass distribution characteristics in each region, calculate the mean and standard deviation of the biomass in the region, and generate the partition estimation parameters:

[0056]

[0057] Among them, μ i is the mean biomass of the ith segmentation area, σ i is the standard deviation, y(x) is the predicted biomass value of pixel x, |R i | is the number of pixels in the region;

[0058] S45. Combine the zoning estimation parameters with the prediction model results to generate a refined grassland biomass zoning estimation map, record the biomass distribution and statistical information of each area, and capture the complex local variation characteristics of the grassland.

[0059] Optionally, the S5 specifically includes:

[0060] S51. Integrate the partition estimation results with multimodal remote sensing data, analyze the error distribution of each data source through the uncertainty propagation model, and use the error distribution function to describe the random characteristics of the error source:

[0061]

[0062] Among them, e(x) is the error component, μ e is the mean error, is the error variance;

[0063] S52. Construct uncertainty propagation formula, calculate uncertainty measure of biomass estimation results of subareas, and describe uncertainty range through error propagation relationship:

[0064]

[0065] Where U(y) is the uncertainty of the output, y is the estimated biomass, and x i is the input feature variable, σ i For variable x i The standard deviation of

[0066] S53. Using uncertainty quantification optimization algorithm, adjust the estimation model parameters based on the optimization objective function:

[0067] min[L(y)+γ·U(y)];

[0068] Where L(y) is the error loss function between the estimated value and the actual observed value, and γ is the uncertainty weight parameter;

[0069] S54, verifying the optimized biomass estimation results, analyzing the mean square error and standardized residual distribution in the region, and verifying whether the uncertainty level of the optimization results meets the predetermined threshold condition;

[0070] S55. Output the optimized grassland biomass estimation result, including the estimated value and its corresponding uncertainty range.

[0071] Optionally, the S6 specifically includes:

[0072] S61. Collect multi-temporal satellite remote sensing data covering the target area to obtain macro-dynamic change information of grassland, and combine it with high-resolution image data obtained by drones to generate remote sensing data sets across time and space scales;

[0073] S62. Perform spatial registration processing on the cross-scale data. Use a multi-scale registration algorithm to align the detailed information of the drone image with the macro coverage data of the satellite image. The registered data is expressed in joint resolution as:

[0074] D(x,t)=Fmerge(Dsatellite(x,t),D UAV (x));

[0075] Among them, D(x,t) is the spatiotemporal remote sensing data, D satellite (x, t) is satellite data, D UAV (x) is the drone image data, F merge is the data fusion function;

[0076] S63. Use the time and space scale compensation algorithm to dynamically balance the time and space resolution D(x,t) in the joint remote sensing data, so that the UAV and satellite data are consistent in the time dimension and complete in the space dimension;

[0077] S64. Combined with the zoning estimation results, the spatiotemporal correlation parameters of grassland biomass are optimized based on the dynamic update model. The updated correlation parameter matrix can be expressed as:

[0078] P(t,x)=α·Pprev(t,x)+β·Pnew(t,x);

[0079] Among them, P(t,x) is the optimized correlation parameter matrix, P prev (t,x) is the previous parameter, P new (t,x) is the correlation parameter of the newly added data, α and β are weight factors;

[0080] S65. Based on the updated spatiotemporal correlation parameters P(t,x), an optimized grassland biomass monitoring result is generated, and a dynamic change trend graph is outputted to characterize the change pattern of grassland biomass across spatiotemporal scales.

[0081] Optionally, the S7 specifically includes:

[0082] S71. Map the optimized grassland biomass monitoring data to a high-resolution spatial grid, and calculate the spatial distribution value of the biomass for each grid cell:

[0083]

[0084] Among them, B(x,y) is the biomass value of the grid unit (x,y), b i (x, y) is the original biomass value of the corresponding unit in the monitoring data, w i is the optimized weight factor;

[0085] S72. Based on the distribution results, a high-resolution grassland biomass spatial distribution map is constructed, and the unobserved areas are filled using an interpolation algorithm to maintain consistency with the observed data;

[0086] S73. Extract the dynamic change trend of grassland biomass in the time dimension, and calculate the overall change rate for the spatial distribution map of each time point:

[0087]

[0088] Among them, R t is the overall rate of change at time t, B t (x i ,y i ) and B t-1 (x i ,y i ) are the biomass values ​​at the current moment and the previous moment respectively;

[0089] S74, statistically analyzing the dynamic change trend data, generating a distribution map of the biomass growth rate and decline rate in the region, and recording the abnormal change area;

[0090] S75. Integrate spatial distribution maps and dynamic change trend analysis results to generate a comprehensive assessment report on grassland ecosystems.

[0091] Optional modules include:

[0092] Remote sensing data acquisition module: collects multi-modal remote sensing data, including high-resolution drone images, multi-temporal satellite remote sensing data and other auxiliary remote sensing data;

[0093] Data preprocessing module: performs radiation correction, geometric correction and noise removal on remote sensing data to generate standardized remote sensing image data sets;

[0094] Feature extraction module: extracts spectral features, texture features and geometric structure features from standardized remote sensing data, and constructs a high-dimensional feature matrix that characterizes the state of grassland ecosystem;

[0095] Dynamic analysis module: Based on the high-dimensional feature matrix, a dynamic spatiotemporal feature analysis framework is designed, and the spatiotemporal correlation parameters of grassland biomass are generated using temporal change decomposition and spatial heterogeneity clustering methods;

[0096] Prediction modeling module: Combine deep learning algorithms and dynamic feature description sets to train nonlinear biomass prediction models to achieve accurate estimation of grassland biomass;

[0097] Zoning Optimization Module: Through the joint analysis of high-resolution images and prediction model results, the heterogeneous grassland areas are stratified using regional segmentation technology to generate refined grassland biomass zoning estimation results;

[0098] Uncertainty analysis module: evaluates the error sources of monitoring data through uncertainty propagation analysis and quantitative optimization algorithms, and generates optimized biomass estimation results;

[0099] Collaborative monitoring module: Through the collaborative monitoring mechanism of UAV and satellite remote sensing, data across time and space scales are integrated, time and space correlation parameters are dynamically updated, and grassland biomass monitoring results are optimized;

[0100] Assessment and visualization module: Generates high-resolution grassland biomass spatial distribution maps and dynamic change trend analysis results, and provides a comprehensive assessment report on grassland ecosystems.

[0101] The beneficial effects of the present invention are:

[0102] The present invention combines multimodal remote sensing data fusion technology, deep learning neural network modeling, dynamic spatiotemporal feature analysis and high-resolution image segmentation technology to achieve accurate estimation and dynamic monitoring of grassland biomass, enabling the system to fully respond to the complex dynamic changes of grassland ecosystems, especially in heterogeneous grassland environments, effectively improving monitoring accuracy and adaptability. At the same time, using uncertainty quantification analysis and optimization algorithms, the uncertainty of multi-source data can be evaluated and optimized, thereby outputting more reliable monitoring results.

[0103] The present invention combines high-resolution and multi-temporal data through the coordinated monitoring mechanism of UAV and satellite remote sensing, dynamically updates the spatiotemporal correlation parameters of grassland biomass, and significantly improves the real-time nature of the monitoring results in the time dimension and the integrity of the spatial dimension. This not only solves the contradiction between high resolution and large-scale coverage in traditional methods, but also makes the generation of spatial distribution maps and dynamic change trends of grassland biomass more accurate and efficient.

[0104] By constructing a dynamic feature description set and an optimized deep learning model, this invention provides a full-process technical framework from feature extraction to model prediction for grassland biomass monitoring, which can provide high-resolution spatial distribution description and time-series dynamic trend analysis of grassland biomass. At the same time, combined with automated regional segmentation and zoning estimation technology, it can capture complex local changes in grasslands, provide comprehensive and detailed grassland ecological assessments, and effectively support scientific decision-making in precision agriculture and grassland ecological management. BRIEF DESCRIPTION OF THE DRAWINGS

[0105] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0106] Figure 1 This is a general framework diagram of a grassland biomass remote sensing monitoring method and system proposed by the present invention;

[0107] Figure 2 The present invention is a flowchart for constructing a grassland biomass prediction model based on dynamic spatiotemporal characteristic analysis. DETAILED DESCRIPTION

[0108] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, which only illustrate the basic structure of the present invention in a schematic manner, and therefore only show the components related to the present invention.

[0109] refer to Figure 1-2 , a remote sensing monitoring method and system for grassland biomass, comprising the following steps:

[0110] S1. Using multimodal remote sensing data fusion technology, we integrate UAV images, high-resolution satellite images, and time-series remote sensing data to generate a cross-resolution, multi-temporal joint feature space, extract grassland spectral features, texture features, and geometric structure features, and construct a high-dimensional feature matrix that characterizes the state of grassland ecosystems.

[0111] S2. Based on the high-dimensional feature matrix and dynamic spatiotemporal feature analysis framework, the temporal change decomposition and spatial heterogeneity clustering methods are used to generate spatiotemporal correlation parameters of grassland biomass and construct a description set of grassland dynamic change characteristics;

[0112] S3. Using the dynamic change feature description set, combined with deep neural network and multi-task optimization algorithm, the nonlinear biomass prediction model is trained to achieve high-precision grassland biomass estimation based on multi-source data characteristics and grassland heterogeneity distribution;

[0113] S4. Combine the prediction model results with high-resolution images, apply regional segmentation technology to perform stratified processing on heterogeneous grassland areas, and generate refined grassland biomass zoning estimation results to capture the complex local changes of grasslands;

[0114] S5. Conduct uncertainty propagation analysis on the partition estimation results and multimodal data, use uncertainty quantification optimization algorithm to evaluate the error sources of monitoring data, and output optimized grassland biomass estimation results;

[0115] S6. Through the coordinated monitoring mechanism of UAV and satellite remote sensing, the zoning estimation results are combined with cross-temporal and spatial scale data, and the spatiotemporal correlation parameters are dynamically updated to optimize the real-time nature of grassland biomass monitoring results in the temporal dimension and the integrity in the spatial dimension;

[0116] S7. Based on the optimized monitoring data, generate high-resolution grassland biomass spatial distribution maps and dynamic change trend analysis results to form a comprehensive assessment of grassland ecosystem status and grassland biomass changes.

[0117] In this implementation, S1 specifically includes:

[0118] S11. Obtain UAV remote sensing image data, collect grassland surface detail information through multispectral sensors, generate high-resolution images, and simultaneously obtain multi-temporal high-resolution satellite image data and other remote sensing data covering the study area;

[0119] S12. Data preprocessing of UAV and satellite images, including radiometric correction, geometric correction and noise removal, to generate standardized image sets and make multi-source data comparable in spectral, geometric and temporal dimensions;

[0120] S13. Apply data registration algorithms to spatially align remote sensing data of different resolutions and multiple temporal phases, and use SIFT algorithm to geometrically register images to generate cross-resolution and multi-temporal remote sensing data sets;

[0121] S14. Extract multimodal features of remote sensing data, including spectral features, texture features and geometric structure features. The spectral features are calculated using the Normalized Difference Vegetation Index (NDVI) and the Enhanced Vegetation Index (EVI):

[0122]

[0123] Among them, NIR represents the reflectivity of the near infrared band, R represents the reflectivity of the red light band, and B represents the reflectivity of the blue light band;

[0124] S15, extracting texture features of the image, using a gray level co-occurrence matrix to calculate contrast, homogeneity and energy indexes of the image, and generating a multi-dimensional texture feature matrix;

[0125] S16, extracting geometric structural features of remote sensing images, combining watershed segmentation algorithm and edge detection algorithm to identify grassland terrain boundaries and morphological structures, and generating a geometric feature description set related to spatial distribution;

[0126] S17. Construct a high-dimensional feature matrix to characterize the state of grassland ecosystem, and fuse the extracted spectral features, texture features and geometric structure features into a joint feature vector.

[0127] In this implementation, S2 specifically includes:

[0128] S21. Using the time series remote sensing data in the high-dimensional feature matrix, a time series decomposition model is established, and the seasonal and trend change characteristics are extracted by the weighted moving average method:

[0129]

[0130] Among them, T(t) is the trend component of the time series, X(ti) is the observation value at the previous i moments, W(i) is the weighting factor, and n is the moving window length;

[0131] S22. Combining the spatial heterogeneity clustering method, the spatial features in the high-dimensional feature matrix are clustered by a density-based spatial clustering algorithm to generate spatial partitions within the grassland area;

[0132] S23. Analyze the dynamic characteristics of each spatial partition and construct a grassland biomass dynamic change model with time and space dimensions as the core:

[0133] M(x,t)=F(S(x),T(t),P(x,t));

[0134] Among them, M(x,t) represents the dynamic characteristics at position x and time t, S(x) is the spatial characteristics, T(t) is the temporal characteristics, and P(x,t) is the external influencing factor;

[0135] S24. Through the adaptive time window method, the spatiotemporal correlation parameters within the partition are dynamically updated, and a grassland dynamic feature description set is generated according to the seasonal fluctuations and heterogeneous distribution changes of grassland biomass.

[0136] In this implementation, S3 specifically includes:

[0137] S31, dividing the dynamically changing feature description set into a training data set and a validation data set, and mapping the spatiotemporal correlation parameters and external environmental variables to a unified numerical range through normalization processing to adapt to the input requirements of the deep neural network;

[0138] S32. Construct a multi-layer deep neural network structure, including an input layer, a hidden layer and an output layer, wherein the input layer receives the key features in the dynamic change feature description set, the hidden layer realizes nonlinear feature conversion through an activation function, and the output layer generates a grassland biomass prediction value:

[0139] y=f(W3·g(W2·g(W1·x+b1)+b2)+b3);

[0140] Where x is the input feature vector, W1, W2, W3 are weight matrices, b1, b2, b3 are bias vectors, g(·) is the activation function, and f(·) is the activation function of the output layer;

[0141] S33. Combined with the multi-task optimization algorithm, the loss function of the grassland biomass prediction task is defined. The loss function consists of the mean square error and the uncertainty quantization error:

[0142] L = α·MSE + β·UQE;

[0143] Among them, α and β are task weight parameters, UQE is the uncertainty measure of the predicted value;

[0144] S34, using the back propagation algorithm to optimize the network weights and bias parameters, adjusting the network structure according to the gradient update principle of the loss function, and gradually minimizing the prediction error;

[0145] S35, verifying the optimized model, and using the validation data set to calculate the accuracy indicators of the prediction results, including correlation coefficients and relative errors, to verify the applicability of the model under grassland heterogeneity distribution and multi-source data characteristics;

[0146] S36. The trained and verified deep neural network model is used as the final nonlinear biomass prediction model to generate the spatial distribution estimation results of grassland biomass.

[0147] In this implementation, S4 specifically includes:

[0148] S41. Perform spatial overlap analysis on the grassland biomass distribution results generated by the prediction model and the high-resolution remote sensing images to calibrate the spatial distribution consistency between the prediction results and the images;

[0149] S42. Based on the calibrated data, the heterogeneous grassland area is segmented using the regional segmentation technology, and the watershed algorithm is used to generate the initial segmentation area according to the spectral difference of the image and the terrain change characteristics;

[0150] S43, optimizing the initial segmented area, combining edge detection method and morphological operation to eliminate over-segmented and under-segmented areas, and the optimized segmented area is expressed in vector form as:

[0151] Ri={x|x∈Ω,g(x)>τi};

[0152] Among them, R i is the i-th segmentation region, x is a pixel point, Ω is the global research area, g(x) is the pixel feature function in the region, τ i is the segmentation threshold;

[0153] S44. Combine the segmented regions, analyze the biomass distribution characteristics in each region, calculate the mean and standard deviation of the biomass in the region, and generate the partition estimation parameters:

[0154]

[0155] Among them, μ i is the mean biomass of the ith segmentation area, σ i is the standard deviation, y(x) is the predicted biomass value of pixel x, |R i | is the number of pixels in the region;

[0156] S45. Combine the zoning estimation parameters with the prediction model results to generate a refined grassland biomass zoning estimation map, record the biomass distribution and statistical information of each area, and capture the complex local variation characteristics of the grassland.

[0157] In this implementation manner, S5 specifically includes:

[0158] S51. Integrate the partition estimation results with multimodal remote sensing data, analyze the error distribution of each data source through the uncertainty propagation model, and use the error distribution function to describe the random characteristics of the error source:

[0159]

[0160] Among them, e(x) is the error component, μ e is the mean error, is the error variance;

[0161] S52. Construct uncertainty propagation formula, calculate uncertainty measure of biomass estimation results of subareas, and describe uncertainty range through error propagation relationship:

[0162]

[0163] Where U(y) is the uncertainty of the output, y is the estimated biomass, and x i is the input feature variable, σ i For variable x i The standard deviation of

[0164] S53. Using uncertainty quantification optimization algorithm, adjust the estimation model parameters based on the optimization objective function:

[0165] min[L(y)+γ·U(y)];

[0166] Where L(y) is the error loss function between the estimated value and the actual observed value, and γ is the uncertainty weight parameter;

[0167] S54, verifying the optimized biomass estimation results, analyzing the mean square error and standardized residual distribution in the region, and verifying whether the uncertainty level of the optimization results meets the predetermined threshold condition;

[0168] S55. Output the optimized grassland biomass estimation result, including the estimated value and its corresponding uncertainty range.

[0169] In this implementation manner, S6 specifically includes:

[0170] S61. Collect multi-temporal satellite remote sensing data covering the target area to obtain macro-dynamic change information of grassland, and combine it with high-resolution image data obtained by drones to generate remote sensing data sets across time and space scales;

[0171] S62. Perform spatial registration processing on the cross-scale data. Use a multi-scale registration algorithm to align the detailed information of the drone image with the macro coverage data of the satellite image. The registered data is expressed in joint resolution as:

[0172] D(x,t)=Fmerge(Dsatellite(x,t),D UAV (x));

[0173] Among them, D(x,t) is the spatiotemporal remote sensing data, D satellite (x, t) is satellite data, D UAV (x) is the drone image data, F merge is the data fusion function;

[0174] S63. Use the time and space scale compensation algorithm to dynamically balance the time and space resolution D(x,t) in the joint remote sensing data, so that the UAV and satellite data are consistent in the time dimension and complete in the space dimension;

[0175] S64. Combined with the zoning estimation results, the spatiotemporal correlation parameters of grassland biomass are optimized based on the dynamic update model. The updated correlation parameter matrix can be expressed as:

[0176] P(t,x)=α·Pprev(t,x)+β·Pnew(t,x);

[0177] Among them, P(t,x) is the optimized correlation parameter matrix, P prev (t,x) is the previous parameter, P new (t,x) is the correlation parameter of the newly added data, α and β are weight factors;

[0178] S65. Based on the updated spatiotemporal correlation parameters P(t,x), an optimized grassland biomass monitoring result is generated, and a dynamic change trend graph is outputted to characterize the change pattern of grassland biomass across spatiotemporal scales.

[0179] In this implementation manner, the S7 specifically includes:

[0180] S71. Map the optimized grassland biomass monitoring data to a high-resolution spatial grid, and calculate the spatial distribution value of the biomass for each grid cell:

[0181]

[0182] Among them, B(x,y) is the biomass value of the grid unit (x,y), b i (x, y) is the original biomass value of the corresponding unit in the monitoring data, w i is the optimized weight factor;

[0183] S72. Based on the distribution results, a high-resolution grassland biomass spatial distribution map is constructed, and the unobserved areas are filled using an interpolation algorithm to maintain consistency with the observed data;

[0184] S73. Extract the dynamic change trend of grassland biomass in the time dimension, and calculate the overall change rate for the spatial distribution map of each time point:

[0185]

[0186] Among them, R t is the overall rate of change at time t, B t (x i ,y i ) and B t-1 (x i ,y i ) are the biomass values ​​at the current moment and the previous moment respectively;

[0187] S74, statistically analyzing the dynamic change trend data, generating a distribution map of the biomass growth rate and decline rate in the region, and recording the abnormal change area;

[0188] S75. Integrate spatial distribution maps and dynamic change trend analysis results to generate a comprehensive assessment report on grassland ecosystems.

[0189] In this embodiment, the following modules are included:

[0190] Remote sensing data acquisition module: collects multi-modal remote sensing data, including high-resolution drone images, multi-temporal satellite remote sensing data and other auxiliary remote sensing data;

[0191] Data preprocessing module: performs radiation correction, geometric correction and noise removal on remote sensing data to generate standardized remote sensing image data sets;

[0192] Feature extraction module: extracts spectral features, texture features and geometric structure features from standardized remote sensing data, and constructs a high-dimensional feature matrix that characterizes the state of grassland ecosystem;

[0193] Dynamic analysis module: Based on the high-dimensional feature matrix, a dynamic spatiotemporal feature analysis framework is designed, and the spatiotemporal correlation parameters of grassland biomass are generated using temporal change decomposition and spatial heterogeneity clustering methods;

[0194] Prediction modeling module: Combine deep learning algorithms and dynamic feature description sets to train nonlinear biomass prediction models to achieve accurate estimation of grassland biomass;

[0195] Zoning Optimization Module: Through the joint analysis of high-resolution images and prediction model results, the heterogeneous grassland areas are stratified using regional segmentation technology to generate refined grassland biomass zoning estimation results;

[0196] Uncertainty analysis module: evaluates the error sources of monitoring data through uncertainty propagation analysis and quantitative optimization algorithms, and generates optimized biomass estimation results;

[0197] Collaborative monitoring module: Through the collaborative monitoring mechanism of UAV and satellite remote sensing, data across time and space scales are integrated, time and space correlation parameters are dynamically updated, and grassland biomass monitoring results are optimized;

[0198] Assessment and visualization module: Generates high-resolution grassland biomass spatial distribution maps and dynamic change trend analysis results, and provides a comprehensive assessment report on grassland ecosystems.

[0199] Embodiment 1:

[0200] In order to verify the feasibility of the present invention in implementation, the present invention was applied to a grassland area in Hainan Tibetan Autonomous Prefecture, Qinghai Province. In order to solve the problems of insufficient data fusion capability, low model accuracy and insufficient response to dynamic changes of grassland in traditional grassland biomass monitoring, a grassland biomass remote sensing monitoring method and system proposed in the present invention were used to dynamically monitor and evaluate the study area. This monitoring combined multimodal remote sensing data, including high-resolution UAV images and multi-phase satellite remote sensing data, and achieved accurate monitoring of grassland ecosystems through deep learning model optimization, biomass dynamic change analysis and spatial distribution evaluation technology.

[0201] In this scenario, multimodal remote sensing data is first collected for the study area. The drone image data was collected in early July 2023, covering a total area of ​​120 square kilometers in the study area, with a resolution of 0.1 meters, mainly recording detailed features of the grassland, such as vegetation coverage and regional texture features. At the same time, multi-phase satellite remote sensing images from May to August 2023 were obtained with a resolution of 10 meters to record the macroscopic changes of the grassland. All data were processed through spatial registration and dynamic spatiotemporal scale compensation technology to generate cross-resolution and cross-phase remote sensing data sets. Through the dynamic spatiotemporal feature analysis module of the present invention, a dynamic change model of grassland biomass was constructed, and key spatiotemporal correlation parameters were extracted to provide basic data for subsequent model optimization and monitoring results analysis.

[0202] After data processing, a deep learning neural network was used to estimate grassland biomass. The model input features included spectral features, texture features, and geometric structure features, totaling 12,000 sample points. After 50 rounds of training, the model had a mean square error of 4.2 g / m on the validation set. 2 , the correlation coefficient reaches 0.92. By comparing with the traditional linear regression model, the accuracy of the model of the present invention is improved by about 25%, which verifies the effectiveness of combining dynamic feature modeling with deep learning.

[0203] Based on the partition optimization technology, the grassland area was divided into detailed partitions according to the grassland heterogeneity and terrain characteristics by taking advantage of the high resolution of drone images. The partition results show that the study area can be divided into 8 main partitions, including dense grassland areas, sparse areas and mixed terrain. The mean distribution of partition biomass generated by regional segmentation technology is as follows. In the dynamic trend analysis, the change of grassland biomass was analyzed in combination with the spatiotemporal renewal model. It was found that the grassland biomass gradually increased from May to July and reached a peak in August. The dynamic change trend in the study area shows that the biomass in the high coverage area increased rapidly, while the biomass in the sparse area changed less.

[0204] Table 1 Distribution table of mean biomass in different regions (unit: g / m 2 )

[0205]

[0206]

[0207] Table 2 Dynamic change trend of grassland biomass (unit: g / m 2 )

[0208] time Average biomass Biomass growth rate (%) May 2023 240.3 - June 2023 275.6 14.7 July 2023 320.8 16.4 August 2023 345.2 7.6

[0209] From the above table, we can see that this study finally generated a high-resolution spatial distribution map of grassland biomass and a dynamic change trend map, and conducted a comprehensive assessment of the grassland ecosystem status. The results show that the method of the present invention can effectively solve the contradiction between spatial resolution and coverage in traditional monitoring methods, while improving the accuracy and dynamic response capability of biomass estimation. In the actual application of the study area, the comprehensive evaluation results of the system have provided key data support for grassland ecological protection and precision agriculture.

[0210] In summary, the present invention successfully solves the limitations of traditional grassland biomass remote sensing monitoring methods in terms of single data, low model accuracy, and insufficient dynamic response by introducing multimodal remote sensing data fusion, dynamic spatiotemporal feature analysis, deep learning modeling, and partition optimization technology. It effectively improves the accuracy of grassland biomass monitoring, dynamically captures the changing trend of grassland biomass, and realizes refined analysis of high-resolution spatial distribution.

[0211] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A remote sensing monitoring method for grassland biomass, characterized in that: The steps include: S1. Using multimodal remote sensing data fusion technology, integrating UAV images, satellite images and time-series remote sensing data, generating a cross-resolution, multi-temporal joint feature space, extracting grassland spectral features, texture features and geometric structure features, and constructing a high-dimensional feature matrix that characterizes the state of grassland ecosystems; S2. Based on the high-dimensional feature matrix and dynamic spatiotemporal feature analysis framework, the temporal change decomposition and spatial heterogeneity clustering methods are used to generate spatiotemporal correlation parameters of grassland biomass and construct a description set of grassland dynamic change characteristics; S3. Using the dynamic change feature description set, combined with deep neural network and multi-task optimization algorithm, the nonlinear biomass prediction model is trained to achieve high-precision grassland biomass estimation based on multi-source data characteristics and grassland heterogeneity distribution; S4. Combine the prediction model results with the images, apply regional segmentation technology to perform stratified processing on heterogeneous grassland areas, and generate refined grassland biomass zoning estimation results to capture the complex local changes of grasslands; S5. Conduct uncertainty propagation analysis on the partition estimation results and multimodal data, use uncertainty quantification optimization algorithm to evaluate the error sources of monitoring data, and output optimized grassland biomass estimation results; S6. Through the coordinated monitoring mechanism of UAV and satellite remote sensing, the zoning estimation results are combined with cross-temporal and spatial scale data, and the spatiotemporal correlation parameters are dynamically updated to optimize the real-time nature of grassland biomass monitoring results in the temporal dimension and the integrity in the spatial dimension; S7. Generate grassland biomass spatial distribution map and dynamic change trend analysis results based on optimized monitoring data to form a comprehensive assessment of grassland ecosystem status and grassland biomass changes.

2. A remote sensing monitoring method for grassland biomass according to claim 1, characterized in that: The S1 specifically includes: S11. Obtain UAV remote sensing image data, collect grassland surface detail information through multispectral sensors, generate images, and simultaneously obtain multi-temporal satellite image data and other remote sensing data covering the study area; S12. Data preprocessing of UAV and satellite images, including radiometric correction, geometric correction and noise removal, to generate standardized image sets and make multi-source data comparable in spectral, geometric and temporal dimensions; S13. Apply data registration algorithms to spatially align remote sensing data of different resolutions and multiple temporal phases, and use SIFT algorithm to geometrically register images to generate cross-resolution and multi-temporal remote sensing data sets; S14. Extract multimodal features of remote sensing data, including spectral features, texture features and geometric structure features. The spectral features are calculated using the Normalized Difference Vegetation Index (NDVI) and the Enhanced Vegetation Index (EVI): Among them, NIR represents the reflectivity of the near infrared band, R represents the reflectivity of the red light band, and B represents the reflectivity of the blue light band; S15, extracting texture features of the image, using a gray level co-occurrence matrix to calculate contrast, homogeneity and energy indexes of the image, and generating a multi-dimensional texture feature matrix; S16, extracting geometric structural features of remote sensing images, combining watershed segmentation algorithm and edge detection algorithm to identify grassland terrain boundaries and morphological structures, and generating a geometric feature description set related to spatial distribution; S17. Construct a high-dimensional feature matrix to characterize the state of grassland ecosystem, and fuse the extracted spectral features, texture features and geometric structure features into a joint feature vector.

3. The remote sensing monitoring method for grassland biomass according to claim 1, characterized in that: The S2 specifically includes: S21. Using the time series remote sensing data in the high-dimensional feature matrix, a time series decomposition model is established, and the seasonal and trend change characteristics are extracted by the weighted moving average method: Among them, T(t) is the trend component of the time series, X(ti) is the observation value at the previous i moments, W(i) is the weighting factor, and n is the moving window length; S22. Combining the spatial heterogeneity clustering method, the spatial features in the high-dimensional feature matrix are clustered by a density-based spatial clustering algorithm to generate spatial partitions within the grassland area; S23. Analyze the dynamic characteristics of each spatial partition and construct a grassland biomass dynamic change model with time and space dimensions as the core: M(x,t)=F(S(x),T(t),P(x,t)); Among them, M(x,t) represents the dynamic characteristics at position x and time t, S(x) is the spatial characteristics, T(t) is the temporal characteristics, and P(x,t) is the external influencing factor; S24. Through the adaptive time window method, the spatiotemporal correlation parameters within the partition are dynamically updated, and a grassland dynamic feature description set is generated according to the seasonal fluctuations and heterogeneous distribution changes of grassland biomass.

4. The remote sensing monitoring method for grassland biomass according to claim 1, characterized in that: The S3 specifically includes: S31, dividing the dynamically changing feature description set into a training data set and a validation data set, and mapping the spatiotemporal correlation parameters and external environmental variables to a unified numerical range through normalization processing to adapt to the input requirements of the deep neural network; S32. Construct a multi-layer deep neural network structure, including an input layer, a hidden layer and an output layer, wherein the input layer receives the key features in the dynamic change feature description set, the hidden layer realizes nonlinear feature conversion through an activation function, and the output layer generates a grassland biomass prediction value: y=f(W3·g(W2·g(W1·x+b1)+b2)+b3); Where x is the input feature vector, W1, W2, W3 are weight matrices, b1, b2, b3 are bias vectors, g(·) is the activation function, and f(·) is the activation function of the output layer; S33. Combined with the multi-task optimization algorithm, the loss function of the grassland biomass prediction task is defined. The loss function consists of the mean square error and the uncertainty quantization error: L = α·MSE + β·UQE; Among them, α and β are task weight parameters, UQE is the uncertainty measure of the predicted value; S34, using the back propagation algorithm to optimize the network weights and bias parameters, adjusting the network structure according to the gradient update principle of the loss function, and gradually minimizing the prediction error; S35, verifying the optimized model, and using the validation data set to calculate the accuracy indicators of the prediction results, including correlation coefficients and relative errors, to verify the applicability of the model under grassland heterogeneity distribution and multi-source data characteristics; S36. The trained and verified deep neural network model is used as the final nonlinear biomass prediction model to generate the spatial distribution estimation results of grassland biomass.

5. The remote sensing monitoring method for grassland biomass according to claim 1, characterized in that: The S4 specifically includes: S41, performing spatial overlap analysis on the grassland biomass distribution results generated by the prediction model and the remote sensing images, and calibrating the spatial distribution consistency between the prediction results and the images; S42. Based on the calibrated data, the heterogeneous grassland area is segmented using the regional segmentation technology, and the watershed algorithm is used to generate the initial segmentation area according to the spectral difference and terrain change characteristics of the image; S43, optimizing the initial segmented area, combining edge detection method and morphological operation to eliminate over-segmented and under-segmented areas, and the optimized segmented area is expressed in vector form as: Ri={x|x∈Ω,g(x)>τi}; Among them, R i is the i-th segmentation region, x is a pixel point, Ω is the global research area, g(x) is the pixel feature function in the region, τ i is the segmentation threshold; S44. Combine the segmented regions, analyze the biomass distribution characteristics in each region, calculate the mean and standard deviation of the biomass in the region, and generate the partition estimation parameters: Among them, μ i is the mean biomass of the ith segmentation area, σ i is the standard deviation, y(x) is the predicted biomass value of pixel x, |R i | is the number of pixels in the region; S45. Combine the zoning estimation parameters with the prediction model results to generate a refined grassland biomass zoning estimation map, record the biomass distribution and statistical information of each area, and capture the complex local variation characteristics of the grassland.

6. The remote sensing monitoring method for grassland biomass according to claim 1, characterized in that: The S5 specifically includes: S51. Integrate the partition estimation results with multimodal remote sensing data, analyze the error distribution of each data source through the uncertainty propagation model, and use the error distribution function to describe the random characteristics of the error source: e(x)~N(μe,σe 2 ); Among them, e(x) is the error component, μ e is the mean error, σ e 2 is the error variance; S52. Construct uncertainty propagation formula, calculate uncertainty measure of biomass estimation results of subareas, and describe uncertainty range through error propagation relationship: Where U(y) is the uncertainty of the output, y is the estimated biomass, and x i is the input feature variable, σ i For variable x i The standard deviation of S53. Using uncertainty quantification optimization algorithm, adjust the estimation model parameters based on the optimization objective function: min[L(y)+γ·U(y)]; Where L(y) is the error loss function between the estimated value and the actual observed value, and γ is the uncertainty weight parameter; S54, verifying the optimized biomass estimation results, analyzing the mean square error and standardized residual distribution in the region, and verifying whether the uncertainty level of the optimization results meets the predetermined threshold condition; S55. Output the optimized grassland biomass estimation result, including the estimated value and its corresponding uncertainty range.

7. The remote sensing monitoring method for grassland biomass according to claim 1, characterized in that: The S6 specifically includes: S61. Collect multi-temporal satellite remote sensing data covering the target area to obtain macro-dynamic change information of grassland, and combine it with image data obtained by drones to generate remote sensing data sets across time and space scales; S62. Perform spatial registration processing on the cross-scale data. Use a multi-scale registration algorithm to align the detailed information of the drone image with the macro coverage data of the satellite image. The registered data is expressed in joint resolution as: D(x,t)=Fmerge(Dsatellite(x,t),D UAV (x)); Among them, D(x,t) is the spatiotemporal remote sensing data, D satellite (x, t) is satellite data, D UAV (x) is the drone image data, F merge is the data fusion function; S63. Use the time and space scale compensation algorithm to dynamically balance the time and space resolution D(x,t) in the joint remote sensing data to ensure the consistency of the UAV and satellite data in the time dimension and the integrity in the space dimension; S64. Combined with the zoning estimation results, the spatiotemporal correlation parameters of grassland biomass are optimized based on the dynamic update model. The updated correlation parameter matrix can be expressed as: P(t,x)=α·Pprev(t,x)+β·Pnew(t,x); Among them, P(t,x) is the optimized correlation parameter matrix, P prev (t,x) is the previous parameter, P new (t,x) is the correlation parameter of the newly added data, α and β are weight factors; S65. Based on the updated spatiotemporal correlation parameters P(t,x), an optimized grassland biomass monitoring result is generated, and a dynamic change trend graph is outputted to characterize the change pattern of grassland biomass across spatiotemporal scales.

8. The remote sensing monitoring method for grassland biomass according to claim 1, characterized in that: The S7 specifically includes: S71, mapping the optimized grassland biomass monitoring data to the spatial grid, and calculating the spatial distribution value of the biomass for each grid unit: Among them, B(x,y) is the biomass value of the grid unit (x,y), b i (x, y) is the original biomass value of the corresponding unit in the monitoring data, w i is the optimized weight factor; S72. Based on the distribution results, a grassland biomass spatial distribution map is constructed, and the unobserved areas are filled using an interpolation algorithm to maintain consistency with the observed data; S73. Extract the dynamic change trend of grassland biomass in the time dimension, and calculate the overall change rate for the spatial distribution map of each time point: Among them, R t is the overall rate of change at time t, B t (x i ,y i ) and B t-1 (x i ,y i ) are the biomass values ​​at the current moment and the previous moment respectively; S74, statistically analyzing the dynamic change trend data, generating a distribution map of the biomass growth rate and decline rate in the region, and recording the abnormal change area; S75. Integrate spatial distribution maps and dynamic change trend analysis results to generate a comprehensive assessment report of grassland ecosystems.

9. A remote sensing monitoring system for grassland biomass, characterized in that: Includes the following modules: Remote sensing data acquisition module: collects multi-modal remote sensing data, including drone images, multi-temporal satellite remote sensing data and other auxiliary remote sensing data; Data preprocessing module: performs radiation correction, geometric correction and noise removal on remote sensing data to generate standardized remote sensing image data sets; Feature extraction module: extracts spectral features, texture features and geometric structure features from standardized remote sensing data, and constructs a high-dimensional feature matrix that characterizes the state of grassland ecosystem; Dynamic analysis module: Based on the high-dimensional feature matrix, a dynamic spatiotemporal feature analysis framework is designed, and the spatiotemporal correlation parameters of grassland biomass are generated using temporal change decomposition and spatial heterogeneity clustering methods; Prediction modeling module: Combine deep learning algorithms and dynamic feature description sets to train nonlinear biomass prediction models to achieve accurate estimation of grassland biomass; Zoning Optimization Module: Through the joint analysis of images and prediction model results, the heterogeneous grassland areas are stratified using regional segmentation technology to generate refined grassland biomass zoning estimation results; Uncertainty analysis module: evaluates the error sources of monitoring data through uncertainty propagation analysis and quantitative optimization algorithms, and generates optimized biomass estimation results; Collaborative monitoring module: Through the collaborative monitoring mechanism of UAV and satellite remote sensing, data across time and space scales are integrated, time and space correlation parameters are dynamically updated, and grassland biomass monitoring results are optimized; Assessment and visualization module: Generates grassland biomass spatial distribution map and dynamic change trend analysis results, and provides a comprehensive assessment report on grassland ecosystems.

Images