Forest biomass remote sensing intelligent estimation method and system

Abstract

The invention discloses a forest biomass remote sensing intelligent estimation method and system. The method comprises the following steps: acquiring multi-source remote sensing data; preprocessing the multi-source remote sensing data to obtain preprocessed image data; acquiring sample data obtained by field investigation; extracting variable data corresponding to the sample data from the preprocessed image data by using a multi-valued extraction-to-point tool; constructing a forest biomass estimation model based on a random forest algorithm, and matching sample data with the extracted variable data to construct a training data set; setting parameters of a forest biomass estimation model, and training the forest biomass estimation model by using the training data set to obtain a trained model; and performing biomass estimation on the image data in the research area by using the trained model to obtain forest biomass data. According to the invention, the precision and efficiency of forest biomass estimation can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of forest data management, and in particular to a remote sensing intelligent estimation method and system for forest biomass. Background Art

[0002] Forest biomass, as an important indicator for measuring the health status and carbon sequestration capacity of forest ecosystems, is of great significance for global climate change research, forest resource management, and sustainable development.

[0003] However, traditional forest biomass estimation methods often rely on ground surveys, which have the disadvantages of long time consumption, high cost, and limited coverage.

[0004] With the rapid development of remote sensing technology, especially the wide application of high-resolution satellite images, new technical approaches have been provided for the rapid and large-area estimation of forest biomass. However, the existing remote sensing estimation methods still face the problems of low accuracy and low efficiency. Summary of the Invention

[0005] Based on this, the present invention provides a remote sensing intelligent estimation method and system for forest biomass to improve the accuracy and efficiency of forest biomass estimation.

[0006] To achieve the above object, the present invention provides a remote sensing intelligent estimation method for forest biomass, and the method includes:

[0007] Obtain multi-source remote sensing data, where the multi-source remote sensing data includes Sentinel-2 optical image data and Sentinel-1 radar image data. The Sentinel-2 optical image data is used to extract optical features, and the Sentinel-1 radar image data is used to extract radar features;

[0008] Preprocess the Sentinel-2 optical image data and the Sentinel-1 radar image data to obtain preprocessed image data;

[0009] Obtain sample data obtained through field surveys, where the sample data includes tree species, diameter at breast height, and tree height;

[0010] Use a multi-value extraction to point tool to extract variable data corresponding to the sample data from the preprocessed image data. The variable data includes biophysical variables, vegetation indices, and texture features;

[0011] Construct a forest biomass estimation model based on the random forest algorithm, and match the sample data with the extracted variable data to construct a training data set;

[0012] Set the parameters of the forest biomass estimation model, and use the training data set to train the forest biomass estimation model to obtain a trained model;

[0013] Using the trained model to estimate the biomass of the image data in the study area to obtain forest biomass data.

[0014] Preferably, the steps of preprocessing Sentinel-2 optical image data and Sentinel-1 radar image data include:

[0015] Using SNAP software to perform 1C to 2A preprocessing on Sentinel-2 image data to generate L2A level products;

[0016] Using SNAP software to perform radiometric calibration and geometric calibration on Sentinel-1 image data.

[0017] Preferably, the steps of using SNAP software to perform 1C to 2A preprocessing on Sentinel-2 image data to generate L2A level products include:

[0018] Applying the Fmask algorithm to generate a cloud mask to remove the cloud-covered areas in the image;

[0019] Resampling the image to unify the resolution to 10 meters;

[0020] Extracting the biophysical variables of the image, including leaf area index and fraction of absorbed photosynthetically active radiation, and calculating using the Biopyhsical Processor module in SNAP software;

[0021] Using the GDAL library to calculate various vegetation indices, including ratio vegetation index and normalized vegetation index;

[0022] Using ENVI software to extract the texture features of the image, adopting the gray-level co-occurrence matrix method, setting the texture window size to 3×3, and obtaining 8 types of texture feature values.

[0023] Preferably, in the step of obtaining the sample data from field surveys, the sample biomass is calculated using the following formula:

[0024] B = 0.12703 * (d 2 * h) 0.79775

[0025] where B is the biomass, d is the diameter at breast height, and h is the tree height.

[0026] Preferably, in the step of setting the parameters of the forest biomass estimation model, the parameters of the forest biomass estimation model include the number of decision trees, the maximum depth, etc., and the parameter setting is optimized by grid search or random search.

[0027] Preferably, the method further includes:

[0028] Using GIS technology, the estimated forest biomass data is subjected to spatial mapping and visual display, and ArcGIS and QGIS are used for spatial mapping.

[0029] To achieve the above object, the present invention also provides a remote sensing intelligent estimation system for forest biomass, and the system includes:

[0030] A first acquisition module, configured to acquire multi-source remote sensing data, where the multi-source remote sensing data includes Sentinel-2 optical image data and Sentinel-1 radar image data, the Sentinel-2 optical image data is used to extract optical features, and the Sentinel-1 radar image data is used to extract radar features;

[0031] A preprocessing module, configured to preprocess the Sentinel-2 optical image data and the Sentinel-1 radar image data to obtain preprocessed image data;

[0032] A second acquisition module, configured to acquire sample data obtained from field investigations, where the sample data includes tree species, diameter at breast height, and tree height;

[0033] An extraction module, configured to use a multi-value extraction to point tool to extract variable data corresponding to the sample data from the preprocessed image data, where the variable data includes biophysical variables, vegetation indices, and texture features;

[0034] A construction module, configured to construct a forest biomass estimation model based on a random forest algorithm, and match the sample data with the extracted variable data to construct a training data set;

[0035] A training setting module, configured to set parameters of the forest biomass estimation model, and use the training data set to train the forest biomass estimation model to obtain a trained model;

[0036] An estimation module, configured to use the trained model to estimate the biomass of the image data in the study area to obtain forest biomass data.

[0037] Preferably, the preprocessing module is specifically configured to:

[0038] Use SNAP software to perform l C to 2A preprocessing on the Sentinel-2 image data to generate L2A level products;

[0039] Use SNAP software to perform radiometric correction and geometric correction processing on the Sentinel-1 image data.

[0040] Preferably, the preprocessing module is specifically configured to:

[0041] Apply the Fmask algorithm to generate a cloud mask to remove the cloud-covered area in the image;

[0042] Resample the image to unify the resolution to 10 meters;

[0043] Extract the biophysical variables of the image, including leaf area index and fraction of absorbed photosynthetically active radiation, and calculate them using the Biopyhsical Processor module in SNAP software;

[0044] Use the GDAL library to calculate various vegetation indices, including ratio vegetation index and normalized vegetation index;

[0045] Use ENVI software to extract the texture features of the image. Adopt the gray-level co-occurrence matrix method, set the texture window size to 3×3, and obtain 8 types of texture feature values.

[0046] Preferably, the second acquisition module is used to calculate the sample biomass using the following formula:

[0047] B = 0.12703 * (d 2 * h) 0.79775

[0048] where B is the biomass, d is the diameter at breast height, and h is the tree height.

[0049] Preferably, the parameters of the forest biomass estimation model include the number of decision trees, the maximum depth, etc., and the parameter settings are optimized by grid search or random search.

[0050] Preferably, the system further includes:

[0051] A display module, which is used to use GIS technology to perform spatial mapping and visual display of the estimated forest biomass data, and use ArcGIS and QGIS for spatial mapping.

[0052] The above-mentioned present invention provides a method and system for remote sensing intelligent estimation of forest biomass, and has the following beneficial effects:

[0053] 1. High precision: Through the fusion of multi-source remote sensing data and advanced image processing technology, the accuracy and reliability of biomass estimation are improved.

[0054] 2. High efficiency: The automated image preprocessing process and intelligent estimation model significantly reduce the time cost of data processing and improve the estimation efficiency.

[0055] 3. Easy to operate: The method of the present invention is simple to operate and easy to promote and apply, providing strong technical support for forest resource management and ecological environment protection.

[0056] 4. Wide application prospects: The present invention can be widely applied to fields such as forest resource inventory, carbon sink assessment, and ecological protection, and has important scientific value and social significance. Description of the Drawings

[0057] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for the description of the embodiments of the present invention will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings.

[0058] Figure 1 is a schematic flowchart of a forest biomass remote sensing intelligent estimation method according to the first embodiment of the present invention;

[0059] Figure 2 is a structural block diagram of a forest biomass remote sensing intelligent estimation system according to the second embodiment of the present invention. Specific Embodiments

[0060] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. 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.

[0061] Please refer to Figure 1 , the first embodiment of the present invention provides a forest biomass remote sensing intelligent estimation method, including steps S101 to S107:

[0062] S101, obtaining multi-source remote sensing data, where the multi-source remote sensing data includes Sentinel-2 optical image data and Sentinel-1 radar image data. The Sentinel-2 optical image data is used to extract optical features, and the Sentinel-1 radar image data is used to extract radar features.

[0063] Specifically, download Sentinel-2 and Sentinel-1 image data from data sources such as the Copernicus Scientific Data Center (CSDB). The Sentinel-2 optical image data is mainly used to extract optical features, such as vegetation indices and texture features; the Sentinel-1 radar image data is used to extract radar features, such as backscattering coefficients, etc.

[0064] S102, preprocessing the Sentinel-2 optical image data and the Sentinel-1 radar image data to obtain preprocessed image data.

[0065] Among them, the steps of preprocessing the Sentinel-2 optical image data and the Sentinel-1 radar image data include:

[0066] The Sentinel-2 image data is preprocessed from Level 1C to Level 2A using the SNAP software to generate Level 2A products; this step includes radiometric calibration, atmospheric correction, etc. to improve the image quality.

[0067] The Sentinel-1 image data is processed for radiometric calibration and geometric calibration using the SNAP software.

[0068] Specifically, the steps of preprocessing the Sentinel-2 image data from Level 1C to Level 2A using the SNAP software to generate Level 2A products include:

[0069] Apply the Fmask algorithm to generate a cloud mask to remove the cloud-covered areas in the image; the generation of the cloud mask can be achieved through software or by writing code using Python;

[0070] Resample the image to a unified resolution of 10 meters for subsequent processing;

[0071] Extract the biophysical variables of the image, including leaf area index (LAI), fraction of absorbed photosynthetically active radiation (FAPAR), and calculate them using the Biopyhsical Processor module in the SNAP software;

[0072] Use the GDAL library to calculate various vegetation indices, including ratio vegetation index (RVI), normalized difference vegetation index (NDVI). The specific calculation formulas are as follows: RVI = NIR / RED, where NIR is the reflectance of the near-infrared band and RED is the reflectance of the red band;

[0073] Use the ENVI software to extract the texture features of the image. Adopt the gray-level co-occurrence matrix (GLCM) method, set the texture window size to 3×3, and obtain 8 types of texture feature values.

[0074] Specifically, use the SNAP software to preprocess the Sentinel-1 image data, including radiometric calibration, geometric calibration, etc. Extract the texture features of the image, and based on the VH and VV polarization images, obtain 16 texture feature values.

[0075] S103, obtain the sample data obtained from field surveys. The sample data includes tree species, diameter at breast height, and tree height.

[0076] Specifically, conduct field surveys within the study area to obtain sample data, including information such as tree species, diameter at breast height, and tree height.

[0077] Among them, in the step of obtaining the sample data obtained from field surveys, calculate the sample biomass according to the allometric equation. The allometric equation varies according to different tree species. For example, for the allometric growth of Chinese fir, the following formula is used to calculate the sample biomass:

[0078] B = 0.12703 * (d 2 * h) 0.79775

[0079] Where B is the biomass, d is the diameter at breast height, and h is the tree height.

[0080] S104. Using the multi - value extraction to point tool, extract the variable data corresponding to the sample data from the pre - processed image data, and the variable data includes biophysical variables, vegetation indices, and texture features.

[0081] S105. Based on the random forest algorithm, construct a forest biomass estimation model, and match the sample data with the extracted variable data to construct a training data set.

[0082] Specifically, use the random forest algorithm based on Python to construct a forest biomass estimation model. Random forest is an ensemble learning method that improves the stability and accuracy of the model by constructing multiple decision trees and voting or averaging.

[0083] Variable correlation analysis is also required, and variables with greater correlation are selected for modeling. Correlation analysis can use methods such as Pearson correlation coefficient.

[0084] Use the variable importance evaluation function in the random forest algorithm to further screen important variables and optimize the model structure.

[0085] S106. Set the parameters of the forest biomass estimation model, and use the training data set to train the forest biomass estimation model to obtain a trained model.

[0086] Among them, the parameters of the forest biomass estimation model include the number of decision trees, maximum depth, etc., and the parameter setting is optimized through grid search or random search.

[0087] S107. Use the trained model to estimate the biomass of the image data in the study area to obtain forest biomass data.

[0088] Using the trained random forest model to estimate the biomass of the image data in the study area can obtain the biomass value of each pixel.

[0089] In addition, as a specific example, the method further includes:

[0090] Using GIS technology, conduct spatial mapping and visual display of the estimated forest biomass data, and use ArcGIS, QGIS for spatial mapping.

[0091] In addition, the accuracy of the estimation results can be verified by calculating evaluation indicators such as the root mean square error (RMSE) and the coefficient of determination (R2). The estimation results are compared with the field survey data to evaluate the accuracy and reliability of the model.

[0092] The present invention has the following technical features:

[0093] 1. Multi-source remote sensing data fusion: The present invention integrates multi-source remote sensing data such as Sentinel-2 optical images and Sentinel-1 radar images, and utilizes the complementary advantages of different data sources to improve the accuracy and reliability of biomass estimation.

[0094] 2. Automated image preprocessing technology: The present invention has developed a set of automated image preprocessing processes, including steps such as image cloud removal, radiometric correction, geometric correction, and texture feature extraction, effectively reducing the complexity and time cost of image preprocessing.

[0095] 3. Intelligent estimation model construction: The present invention uses the random forest algorithm based on Python, combines multi-dimensional variables such as biophysical variables, vegetation indices, and texture features, and constructs an intelligent forest biomass estimation model. Through variable correlation analysis and variable importance evaluation, the model structure is further optimized to improve the estimation accuracy of the model.

[0096] 4. Biomass spatial distribution mapping: The present invention uses GIS technology to perform spatial mapping and visualization of the estimated forest biomass data, providing an intuitive and convenient tool for forest resource management and ecological environment protection.

[0097] The remote sensing intelligent estimation method for forest biomass proposed in this embodiment has the following beneficial effects:

[0098] 1. High accuracy: Through multi-source remote sensing data fusion and advanced image processing technology, the accuracy and reliability of biomass estimation are improved.

[0099] 2. High efficiency: The automated image preprocessing process and intelligent estimation model significantly reduce the time cost of data processing and improve the estimation efficiency.

[0100] 3. Easy to operate: The method of the present invention is simple to operate and easy to promote and apply, providing strong technical support for forest resource management and ecological environment protection.

[0101] 4. Wide application prospects: The present invention can be widely applied to fields such as forest resource inventory, carbon sink assessment, and ecological protection, and has important scientific value and social significance.

[0102] Please refer to Figure 2 , the second embodiment of the present invention provides a remote sensing intelligent estimation system for forest biomass, and the system includes:

[0103] The first acquisition module is used to acquire multi-source remote sensing data, where the multi-source remote sensing data includes Sentinel-2 optical image data and Sentinel-1 radar image data. The Sentinel-2 optical image data is used to extract optical features, and the Sentinel-1 radar image data is used to extract radar features;

[0104] The preprocessing module is used to preprocess the Sentinel-2 optical image data and the Sentinel-1 radar image data to obtain preprocessed image data;

[0105] The second acquisition module is used to acquire sample data obtained from field surveys, where the sample data includes tree species, diameter at breast height, and tree height;

[0106] The extraction module is used to extract variable data corresponding to the sample data from the preprocessed image data by using a multi-value extraction to point tool. The variable data includes biophysical variables, vegetation indices, and texture features;

[0107] The construction module is used to construct a forest biomass estimation model based on the random forest algorithm and match the sample data with the extracted variable data to construct a training data set;

[0108] The setting and training module is used to set the parameters of the forest biomass estimation model and use the training data set to train the forest biomass estimation model to obtain a trained model;

[0109] The estimation module is used to estimate the biomass of the image data in the study area by using the trained model to obtain forest biomass data.

[0110] In this embodiment, the preprocessing module is specifically used for:

[0111] Use SNAP software to perform 1C to 2A preprocessing on the Sentinel-2 image data to generate L2A level products;

[0112] Use SNAP software to perform radiometric correction and geometric correction processing on the Sentinel-1 image data.

[0113] In this embodiment, the preprocessing module is specifically used for:

[0114] Apply the Fmask algorithm to generate a cloud mask to remove the cloud-covered areas in the image;

[0115] Resample the image to unify the resolution to 10 meters;

[0116] Extract the biophysical variables of the image, including leaf area index and fraction of absorbed photosynthetically active radiation, and calculate them using the Biopyhsical Processor module in SNAP software;

[0117] The GDAL library is used to calculate various vegetation indices, including the ratio vegetation index and the normalized vegetation index;

[0118] The ENVI software is used to extract the texture features of the image. The gray-level co-occurrence matrix method is adopted, the texture window size is set to 3×3, and 8 types of texture feature values are obtained.

[0119] In this embodiment, the second acquisition module is used to calculate the sample biomass by the following formula:

[0120] B = 0.12703 * (d 2 * h) 0.79775

[0121] where B is the biomass, d is the diameter at breast height, and h is the tree height.

[0122] In this embodiment, the parameters of the forest biomass estimation model include the number of decision trees, the maximum depth, etc., and the parameter settings are optimized by grid search or random search.

[0123] In this embodiment, the system further includes:

[0124] A display module, which is used to use GIS technology to perform spatial mapping and visual display on the estimated forest biomass data, and use ArcGIS and QGIS for spatial mapping.

[0125] Through the forest biomass remote sensing intelligent estimation system proposed in this embodiment, the following beneficial effects are achieved:

[0126] 1. High precision: Through the fusion of multi-source remote sensing data and advanced image processing technology, the accuracy and reliability of biomass estimation are improved.

[0127] 2. High efficiency: The automated image preprocessing process and intelligent estimation model significantly reduce the time cost of data processing and improve the estimation efficiency.

[0128] 3. Easy to operate: The method of the present invention is simple to operate and easy to promote and apply, providing strong technical support for forest resource management and ecological environment protection.

[0129] 4. Wide application prospects: The present invention can be widely applied to forest resource inventory, carbon sink assessment, ecological protection and other fields, and has important scientific value and social significance.

[0130] The meanings of "first" and "second" in the above-mentioned module / unit are only used to distinguish different modules / units, and are not used to limit which module / unit has a higher priority or other limiting meanings. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or modules does not have to be limited to those steps or modules clearly listed, but may include other steps or modules that are not clearly listed or are inherent to these processes, methods, products or devices. The division of modules in this application is only a logical division, and there may be other division methods in actual implementation.

[0131] For the specific limitations of the forest biomass remote sensing intelligent estimation system, reference can be made to the limitations of the forest biomass remote sensing intelligent estimation method in the above text, which will not be elaborated here. Each module in the above forest biomass remote sensing intelligent estimation system can be implemented in whole or in part by software, hardware and their combination. Each of the above modules can be embedded in the processor of the computer device in hardware form or independent of it, or stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to each of the above modules.

[0132] Those of ordinary skill in the art can understand that all or part of the processes of implementing the methods in the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the above method embodiments. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0133] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the above-mentioned division of each functional unit and module is used as an example. In actual applications, the above functions can be allocated to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0134] The above-described embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit 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 recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A remote sensing intelligent estimation method for forest biomass, characterized in that, The method includes: Obtain multi-source remote sensing data, where the multi-source remote sensing data includes Sentinel-2 optical image data and Sentinel-1 radar image data. The Sentinel-2 optical image data is used to extract optical features, and the Sentinel-1 radar image data is used to extract radar features; Preprocess the Sentinel-2 optical image data and the Sentinel-1 radar image data to obtain preprocessed image data; Obtain sample data obtained from on-site surveys, where the sample data includes tree species, diameter at breast height (DBH), and tree height; Use the multi-value extraction to point tool to extract variable data corresponding to the sample data from the preprocessed image data. The variable data includes biophysical variables, vegetation indices, and texture features; Construct a forest biomass estimation model based on the random forest algorithm, and match the sample data with the extracted variable data to construct a training dataset; Set the parameters of the forest biomass estimation model, and use the training dataset to train the forest biomass estimation model to obtain a trained model; Use the trained model to estimate the biomass of the image data in the study area to obtain forest biomass data.

2. The forest biomass remote sensing intelligent estimation method according to claim 1, wherein The steps for preprocessing the Sentinel-2 optical image data and the Sentinel-1 radar image data include: Perform 1C to 2A preprocessing on the Sentinel-2 image data using SNAP software to generate L2A products; Perform radiometric correction and geometric correction on the Sentinel-1 image data using SNAP software.

3. The forest biomass remote sensing intelligent estimation method according to claim 2, wherein The steps for performing 1C to 2A preprocessing on the Sentinel-2 image data using SNAP software to generate L2A products include: Apply the Fmask algorithm to generate a cloud mask to remove the cloud-covered areas in the image; Resample the image to unify the resolution to 10 meters; Extract the biophysical variables of the image, including leaf area index and fraction of absorbed photosynthetically active radiation, and calculate using the BiopyhsicalProcessor module in SNAP software; Calculate various vegetation indices, including ratio vegetation index and normalized vegetation index, using the GDAL library; Extract the texture features of the image using ENVI software. Using the gray-level co-occurrence matrix method, set the texture window size to 3×3 to obtain 8 types of texture feature values.

4. The forest biomass remote sensing intelligent estimation method according to claim 1, characterized in that In the step of obtaining sample data obtained from on-site surveys, the sample biomass is calculated using the following formula: B = 0.12703 * (d 2 * h) 0.79775 where B is the biomass, d is the diameter at breast height, and h is the tree height.

5. The forest biomass remote sensing intelligent estimation method according to claim 1, characterized in that, In the step of setting the parameters of the forest biomass estimation model, the parameters of the forest biomass estimation model include the number of decision trees, maximum depth, etc. The parameter setting is optimized through grid search or random search.

6. The forest biomass remote sensing intelligent estimation method according to claim 1, characterized in that The method further includes: Using GIS technology, perform spatial mapping and visual display of the estimated forest biomass data, and use ArcGIS and QGIS for spatial mapping.

7. A remote sensing intelligent estimation system for forest biomass, characterized in that, The system includes: A first acquisition module for obtaining multi-source remote sensing data, where the multi-source remote sensing data includes Sentinel-2 optical image data and Sentinel-1 radar image data. The Sentinel-2 optical image data is used to extract optical features, and the Sentinel-1 radar image data is used to extract radar features; A preprocessing module for preprocessing Sentinel-2 optical image data and Sentinel-1 radar image data to obtain preprocessed image data; A second acquisition module for acquiring sample data obtained from on-site surveys, the sample data including tree species, diameter at breast height, and tree height; An extraction module for extracting variable data corresponding to the sample data from the preprocessed image data by using a multi-value extraction to point tool, the variable data including biophysical variables, vegetation indices, and texture features; A construction module for constructing a forest biomass estimation model based on a random forest algorithm and matching the sample data with the extracted variable data to construct a training data set; A setting training module for setting parameters of the forest biomass estimation model and training the forest biomass estimation model by using the training data set to obtain a trained model; An estimation module for estimating the biomass of the image data in the study area by using the trained model to obtain forest biomass data.

8. The forest biomass remote sensing intelligent estimation system according to claim 7, characterized in that The preprocessing module is specifically used for: Performing 1C to 2A preprocessing on the Sentinel-2 image data by using SNAP software to generate L2A products; Performing radiometric correction and geometric correction processing on the Sentinel-1 image data by using SNAP software.

9. The forest biomass remote sensing intelligent estimation system according to claim 8, characterized in that The preprocessing module is specifically used for: Applying the Fmask algorithm to generate a cloud mask to remove cloud-covered areas in the image; Resampling the image to unify the resolution to 10 meters; Extracting biophysical variables of the image, including leaf area index and fraction of absorbed photosynthetically active radiation, and calculating by using the BiopyhsicalProcessor module in SNAP software; Calculating various vegetation indices, including ratio vegetation index and normalized vegetation index, by using the GDAL library; Extracting the texture features of the image by using ENVI software, adopting the gray-level co-occurrence matrix method, setting the texture window size to 3×3, and obtaining 8 types of texture feature values.

10. The forest biomass remote sensing intelligent estimation system according to claim 7, characterized in that The second acquisition module is used to calculate the sample biomass by the following formula: B = 0.12703 * (d 2 * h) 0.79775 where B is the biomass, d is the diameter at breast height, and h is the tree height.

Images