A forest damage area identification method, system, electronic device and storage medium
By preprocessing and calculating feature parameters of Sentinel-1 GRD and SLC data, and combining them with the K-means clustering algorithm, the problem of low efficiency and poor accuracy in forest destruction area identification in existing technologies has been solved, and efficient and accurate automatic identification of forest destruction areas has been achieved.
Patent Information
- Application Number
- CN202310369038.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-04-10
- Publication Date
- 2025-11-21
- Estimated Expiration
- 2043-04-10
AI Technical Summary
Existing methods for identifying deforestation areas based on radar data are inefficient and inaccurate, and lack research on the comprehensive use of backscattering intensity parameters, vegetation indices, and polarization parameters.
By acquiring and preprocessing Sentinel-1 GRD and SLC data, the optimal feature parameters, including backscattering intensity, vegetation index, and polarization parameters, were calculated. The K-means clustering algorithm was then used to identify forest destruction areas.
It improves the efficiency and accuracy of forest destruction area identification, realizes automatic identification of forest destruction areas, avoids the inefficiency and low accuracy of manual monitoring methods, and is suitable for large-scale near real-time monitoring.
Smart Images
Figure CN116403111B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of forest destruction area identification, and in particular to a method, system, electronic device, and storage medium for forest destruction area identification. Background Technology
[0002] Deforestation in tropical regions is considered one of the world's major environmental threats. Synthetic Aperture Radar (SAR) remote sensing plays a crucial role in the continuous monitoring and identification of deforestation areas in cloudy regions because it is unaffected by weather conditions. Currently, most deforestation identification based on radar data relies on time-consuming and labor-intensive manual supervision methods, which limit its application over large spatial areas. Furthermore, manual supervision methods are inefficient and inaccurate in identifying deforestation areas. In addition, backscattering intensity parameters, vegetation indices, and polarization parameters can all effectively characterize ground features and are widely used in forest-related research; however, research that comprehensively considers these three radar characteristics for deforestation area identification remains lacking. Summary of the Invention
[0003] The purpose of this invention is to provide a method, system, electronic device, and storage medium for identifying forest destruction areas, so as to improve the efficiency and accuracy of forest destruction area identification.
[0004] To achieve the above objectives, the present invention provides the following solution:
[0005] A method for identifying deforestation areas includes:
[0006] Acquire Sentinel-1 GRD data and Sentinel-1 SLC data for the forest area to be identified;
[0007] The Sentinel-1 GRD data undergoes a first preprocessing step, and the Sentinel-1 SLC data undergoes a second preprocessing step to obtain preprocessed GRD data and preprocessed SLC data. The first preprocessing step includes orbit correction, radiometric correction, thermal noise and edge noise removal, terrain correction, and filtering. The second preprocessing step includes orbit correction, radiometric correction, multi-view, polarization matrix generation, terrain correction, filtering, and polarization decomposition.
[0008] Based on the preprocessed GRD data and the preprocessed SLC data, the optimal feature parameters for all pixels in the forest region to be identified are calculated. The optimal feature parameters include a first optimal feature parameter, a second optimal feature parameter, and a third optimal feature parameter. The first optimal feature parameter is determined based on the separation of backscattering intensity parameters between the forest destruction area and the environmental background. The backscattering intensity parameters include co-polarized backscattering intensity and cross-polarized backscattering intensity. The second optimal feature parameter is determined based on the separation of vegetation indices between the forest destruction area and the environmental background. The vegetation indices include radar vegetation index, polarimetric radar vegetation index, and radar forest degradation index. The third optimal feature parameter is determined based on the separation of polarization parameters between the forest destruction area and the environmental background. The polarization parameters include the diagonal parameter of the polarization matrix, polarization entropy, polarization anisotropy, and polarization angle.
[0009] Based on the optimal feature parameters of all the pixels, the forest destruction zone of the forest area to be identified is determined using the K-means clustering algorithm.
[0010] Optionally, the number of categories in the K-means clustering algorithm is set to 2, and the number of iterations is set to 25.
[0011] Optionally, based on the preprocessed GRD data and the preprocessed SLC data, the optimal feature parameters for all pixels in the forest region to be identified are calculated, specifically including:
[0012] Based on the preprocessed GRD data, calculate the first optimal feature parameter and the second optimal feature parameter for all pixels in the forest region to be identified.
[0013] Based on the preprocessed SLC data, the third optimal feature parameters of all pixels in the forest region to be identified are calculated.
[0014] Optionally, it also includes:
[0015] A three-dimensional feature space model of the forest region to be identified is established based on the optimal feature parameters of all the pixels; the coordinate axes of the three-dimensional feature space model are the first optimal feature parameter, the second optimal feature parameter, and the third optimal feature parameter, respectively.
[0016] Optionally, the process of determining the optimal feature parameters specifically includes:
[0017] Acquire Sentinel 1 GRD sample data and Sentinel 1 SLC sample data in the forest experimental area;
[0018] The Sentinel 1 GRD sample data is subjected to a first preprocessing, and the Sentinel 1 SLC sample data is subjected to a second preprocessing to obtain preprocessed GRD sample data and preprocessed SLC sample data.
[0019] Based on the preprocessed GRD sample data, calculate the same polarization backscattering intensity, cross polarization backscattering intensity, radar vegetation index, polarization radar vegetation index, and radar forest degradation index for all pixels in the forest experimental area.
[0020] Based on the preprocessed SLC sample data, calculate the diagonal parameters of the polarization matrix, polarization entropy, polarization anisotropy and polarization angle of all pixels in the forest experimental area.
[0021] Calculate the same polarization backscattering intensity, cross polarization backscattering intensity, radar vegetation index, polarization radar vegetation index, radar forest degradation index, polarization matrix diagonal parameter, polarization entropy, polarization anisotropy, and polarization angle of all pixels to determine the separation degree between the forest destruction area and the environmental background in the experimental area.
[0022] Based on all the said separation degrees, a first optimal feature parameter, a second optimal feature parameter, and a third optimal feature parameter are determined.
[0023] Optionally, the following calculations are made for all pixels: the same-polarization backscattering intensity, the cross-polarization backscattering intensity, the radar vegetation index, the polarization radar vegetation index, the radar forest degradation index, the diagonal parameter of the polarization matrix, the polarization entropy, the polarization anisotropy, and the separation degree of the polarization angle between the deforestation area and the environmental background. Specifically, this includes:
[0024] Using formula Calculate the same-polarization backscattering intensity, the cross-polarization backscattering intensity, the radar vegetation index, the polarization radar vegetation index, the radar forest degradation index, the diagonal parameter of the polarization matrix, the polarization entropy, the polarization anisotropy, and the separation degree of the polarization angle between the forest destruction area and the environmental background; where μ1 is the average value of a certain feature parameter of all pixels in the forest destruction area; μ2 is the average value of a certain feature parameter of all pixels in the environmental background; σ1 is the standard deviation of a certain feature parameter of all pixels in the forest destruction area; and σ2 is the standard deviation of a certain feature parameter of all pixels in the environmental background.
[0025] Optionally, based on all the said separation degrees, a first optimal feature parameter, a second optimal feature parameter, and a third optimal feature parameter are determined, specifically including:
[0026] The parameter with the highest separation degree among the backscattering intensity parameters is selected as the first optimal characteristic parameter;
[0027] The parameter with the highest separation degree among the vegetation indices is selected as the second optimal feature parameter.
[0028] Among the polarization parameters, the parameter with the highest separation degree is selected as the third optimal characteristic parameter.
[0029] A deforestation zone identification system, comprising:
[0030] The data acquisition module is used to acquire Sentinel-1 GRD data and Sentinel-1 SLC data of the forest area to be identified;
[0031] The preprocessing module is used to perform a first preprocessing on the Sentinel-1 GRD data and a second preprocessing on the Sentinel-1 SLC data to obtain preprocessed GRD data and preprocessed SLC data. The first preprocessing includes orbit correction, radiometric correction, thermal noise and edge noise removal, terrain correction, and filtering. The second preprocessing includes orbit correction, radiometric correction, multi-view, polarization matrix generation, terrain correction, filtering, and polarization decomposition.
[0032] The feature calculation module is used to calculate the optimal feature parameters for all pixels in the forest area to be identified based on the preprocessed GRD data and the preprocessed SLC data. The optimal feature parameters include a first optimal feature parameter, a second optimal feature parameter, and a third optimal feature parameter. The first optimal feature parameter is determined based on the separation of backscattering intensity parameters between the forest destruction area and the environmental background. The backscattering intensity parameters include co-polarized backscattering intensity and cross-polarized backscattering intensity. The second optimal feature parameter is determined based on the separation of vegetation indices between the forest destruction area and the environmental background. The vegetation indices include radar vegetation index, polarimetric radar vegetation index, and radar forest degradation index. The third optimal feature parameter is determined based on the separation of polarization parameters between the forest destruction area and the environmental background. The polarization parameters include the diagonal parameter of the polarization matrix, polarization entropy, polarization anisotropy, and polarization angle.
[0033] The identification module is used to determine the forest destruction area of the forest area to be identified by using the K-means clustering algorithm based on the best feature parameters of all the pixels.
[0034] An electronic device includes a memory and a processor, wherein the memory stores a computer program, and the processor runs the computer program to enable the electronic device to perform the above-described method for identifying deforestation areas.
[0035] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for identifying deforestation areas.
[0036] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects:
[0037] The forest destruction zone identification method of this invention preprocesses the acquired Sentinel-1 GRD and SLC data of the forest area to be identified, and calculates the optimal feature parameters for all pixels. These optimal feature parameters are determined based on the separation degree of backscattering intensity parameter between the forest destruction zone and the environmental background, the separation degree of vegetation index between the forest destruction zone and the environmental background, and the separation degree of polarization parameter between the forest destruction zone and the environmental background. The optimal feature parameters of all pixels are then clustered using the K-means clustering algorithm to determine the forest destruction zone of the forest area to be identified. This invention replaces manual supervision methods, improving the efficiency and accuracy of forest destruction zone identification. Attached Figure Description
[0038] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0039] Figure 1 Flowchart of the forest destruction area identification method provided by the present invention;
[0040] Figure 2 This is a flowchart of the forest destruction area identification method of the present invention in practical application;
[0041] Figure 3 This is a schematic diagram of the three-dimensional feature space model of the experimental area established in Example 1. Detailed Implementation
[0042] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0043] The purpose of this invention is to provide a method, system, electronic device, and storage medium for identifying forest destruction areas, so as to improve the efficiency and accuracy of forest destruction area identification.
[0044] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.
[0045] like Figure 1 and Figure 2 As shown, the present invention provides a method for identifying deforestation areas, comprising:
[0046] Step 101: Acquire Sentinel-1 GRD (Ground Range Detected) data and Sentinel-1 SLC (Single Look Complex) data of the forest area to be identified.
[0047] Step 102: Perform a first preprocessing on the Sentinel-1 GRD data and a second preprocessing on the Sentinel-1 SLC data to obtain preprocessed GRD data and preprocessed SLC data. The first preprocessing includes orbit correction, radiometric correction, thermal noise and edge noise removal, terrain correction, and filtering; the second preprocessing includes orbit correction, radiometric correction, multi-look, polarization matrix generation, terrain correction, filtering, and polarization decomposition.
[0048] Step 103: Based on the preprocessed GRD data and the preprocessed SLC data, calculate the optimal feature parameters for all pixels in the forest region to be identified; the optimal feature parameters include a first optimal feature parameter, a second optimal feature parameter, and a third optimal feature parameter; the first optimal feature parameter is determined based on the separation of the backscattering intensity parameter between the forest destruction area and the environmental background; the second optimal feature parameter is determined based on the separation of the vegetation index between the forest destruction area and the environmental background; the third optimal feature parameter is determined based on the separation of the polarization parameter between the forest destruction area and the environmental background.
[0049] As an optional implementation, step 103 specifically includes:
[0050] Based on the preprocessed GRD data, calculate the first optimal feature parameter and the second optimal feature parameter for all pixels in the forest region to be identified.
[0051] Based on the preprocessed SLC data, the third optimal feature parameters of all pixels in the forest region to be identified are calculated.
[0052] In practical applications, the process of determining the optimal feature parameters specifically includes:
[0053] Obtain Sentinel-1 GRD and Sentinel-1 SLC sample data from the forest experimental area. In practical applications, search and download the Sentinel-1 GRD and SLC sample data by entering the query date and the latitude and longitude of the forest experimental area on the official Alaska website (https: / / search.asf.alaska.edu / # / ).
[0054] The Sentinel-1 GRD sample data undergoes a first preprocessing step, and the Sentinel-1 SLC sample data undergoes a second preprocessing step, resulting in preprocessed GRD and SLC sample data. In practical applications, SNAP software is used to preprocess the Sentinel-1 GRD and SLC sample data. The preprocessing of the Sentinel-1 GRD sample data includes orbit correction, radiometric correction, thermal and edge noise removal, terrain correction, and filtering. The preprocessing of the Sentinel-1 SLC sample data includes orbit correction, radiometric correction, multi-look processing, polarization matrix generation, terrain correction, filtering, and polarization decomposition. The preprocessed GRD sample data is used to calculate radar backscattering intensity parameters and vegetation indices, while the preprocessed SLC sample data is used to calculate polarization parameters. Among them, the backscattering intensity parameters include co-polarized backscattering intensity and cross-polarized backscattering intensity; the vegetation indices include radar vegetation index, polarimetric radar vegetation index and radar forest degradation index; the polarization parameters include the diagonal parameter of the polarization matrix, polarization entropy, polarization anisotropy and polarization angle.
[0055] Based on the preprocessed GRD sample data, calculate the same polarization backscattering intensity, cross polarization backscattering intensity, radar vegetation index, polarization radar vegetation index, and radar forest degradation index for all pixels in the forest experimental area.
[0056] Based on the preprocessed SLC sample data, the diagonal parameters of the polarization matrix, polarization entropy, polarization anisotropy, and polarization angle of all pixels in the forest experimental area are calculated.
[0057] Calculate the same polarization backscattering intensity, cross polarization backscattering intensity, radar vegetation index, polarization radar vegetation index, radar forest degradation index, polarization matrix diagonal parameter, polarization entropy, polarization anisotropy, and polarization angle of all pixels to determine the separation degree between the forest destruction area and the environmental background in the experimental area.
[0058] Specifically, based on all the characteristic parameters calculated above, the separation index (M-index) is used to analyze the separation between the deforestation area and the environmental background. The formula for calculating the M-index is:
[0059]
[0060] Where μ1 is the average value of a certain feature parameter of all pixels in the forest destruction area; μ2 is the average value of a certain feature parameter of all pixels in the background environment; σ1 is the standard deviation of a certain feature parameter of all pixels in the forest destruction area; and σ2 is the standard deviation of a certain feature parameter of all pixels in the background environment.
[0061] Based on all the said separation degrees, a first optimal feature parameter, a second optimal feature parameter, and a third optimal feature parameter are determined.
[0062] Specifically, the parameter with the highest separation degree among the backscattering intensity parameters is selected as the first optimal characteristic parameter.
[0063] The parameter with the highest separation among the vegetation indices is selected as the second optimal feature parameter.
[0064] Among the polarization parameters, the parameter with the highest separation degree is selected as the third optimal characteristic parameter.
[0065] A three-dimensional feature space model of the forest experimental area is established based on the selected optimal feature parameters, such as... Figure 3 As shown, the deforestation area and the environmental background can be clearly distinguished. The deforestation area is clustered at the bottom of the 3D feature space model, while the environmental background is distributed in the middle and upper parts of the 3D feature space model.
[0066] This invention comprehensively considers three types of characteristic parameters of radar data, which is theoretically superior to the traditional method that only uses radar backscatter intensity. The vegetation index is closely related to forest biomass, and the polarization parameter can characterize physical properties such as forest canopy structure, increasing the distinguishability between forest destruction areas and the environmental background.
[0067] Based on the three optimal feature parameters determined above, a method for identifying forest destruction areas is developed using the K-means clustering algorithm. This step mainly involves applying the K-means clustering algorithm to the constructed three-dimensional feature space model to achieve automatic identification of forest destruction areas. The K-means clustering algorithm is implemented in ENVI software.
[0068] Specifically: (1) Input: Layer stacking of the three best feature parameters of each pixel to generate an RGB image.
[0069] (2) Then, the K-means clustering algorithm is applied to divide the RGB images into two classes.
[0070] Using sample data visually interpreted from Google Earth, the accuracy of the proposed method was quantitatively evaluated and compared with four supervised classifiers. These four supervised classification methods were implemented using ENVI software, with the parameters of the supervised classifiers set to ENVI's default settings. Combining high-resolution Google Earth imagery, 1000 sample points were selected for both the deforestation area and the environmental background in each experimental area. Overall accuracy (OA), Kappa coefficient, and F1 score were used to quantitatively evaluate the recognition results of the proposed method.
[0071] The forest destruction zone identification method of this invention is proposed by coupling a three-dimensional feature space model formed by radar backscattering intensity, vegetation index, and polarization parameters with a clustering algorithm. Tested in three experimental areas, the overall accuracy (OA), Kappa coefficient, and F1 score were 88.1%-98.3%, 0.75-0.96, and 90.2%-98.5%, respectively. Compared with four supervised classifiers (Object-Oriented Classification (OOC), Maximum Likelihood Classification (MLC), Neural Network Classification (NN), and Support Vector Machine (SVM)), the method of this invention achieves accuracy comparable to MLC, NN, and SVM, and achieves better results than OOC, realizing automatic identification of forest destruction zones and avoiding time-consuming and laborious sample selection and manual intervention. Experiments show that this method can accurately and quickly identify forest destruction zones and has the potential to be applied to near real-time monitoring of large-scale forest destruction.
[0072] Step 104: Based on the optimal feature parameters of all the pixels, use the K-means clustering algorithm to determine the forest destruction area of the forest region to be identified. In practical applications, the number of classes in the K-means clustering algorithm is set to 2, and the number of iterations is set to 25.
[0073] Furthermore, the present invention can also establish a three-dimensional feature space model of the forest region to be identified based on the optimal feature parameters of all the pixels; the coordinate axes of the three-dimensional feature space model are the first optimal feature parameter, the second optimal feature parameter, and the third optimal feature parameter, respectively.
[0074] Based on the three-dimensional feature space model, the forest destruction zone of the forest area to be identified is determined.
[0075] This invention also visualizes the optimal feature parameters of the forest area to be identified by establishing a three-dimensional feature space model, which can intuitively show the forest destruction area of the forest area to be identified.
[0076] Example 2
[0077] In order to implement the method corresponding to Embodiment 1 above and achieve the corresponding functions and technical effects, a forest destruction area identification system is provided below, including:
[0078] The data acquisition module is used to acquire Sentinel-1 GRD data and Sentinel-1 SLC data for the forest area to be identified.
[0079] The preprocessing module is used to perform a first preprocessing on the Sentinel-1 GRD data and a second preprocessing on the Sentinel-1 SLC data to obtain preprocessed GRD data and preprocessed SLC data. The first preprocessing includes orbit correction, radiometric correction, thermal noise and edge noise removal, terrain correction, and filtering. The second preprocessing includes orbit correction, radiometric correction, multi-view, polarization matrix generation, terrain correction, filtering, and polarization decomposition.
[0080] The feature calculation module is used to calculate the optimal feature parameters for all pixels in the forest region to be identified based on the preprocessed GRD data and the preprocessed SLC data. The optimal feature parameters include a first optimal feature parameter, a second optimal feature parameter, and a third optimal feature parameter. The first optimal feature parameter is determined based on the separation of backscattering intensity parameters between the forest destruction area and the environmental background. The backscattering intensity parameters include co-polarized backscattering intensity and cross-polarized backscattering intensity. The second optimal feature parameter is determined based on the separation of vegetation indices between the forest destruction area and the environmental background. The vegetation indices include radar vegetation index, polarimetric radar vegetation index, and radar forest degradation index. The third optimal feature parameter is determined based on the separation of polarization parameters between the forest destruction area and the environmental background. The polarization parameters include the diagonal parameter of the polarization matrix, polarization entropy, polarization anisotropy, and polarization angle.
[0081] The identification module is used to determine the forest destruction area of the forest area to be identified by using the K-means clustering algorithm based on the best feature parameters of all the pixels.
[0082] Example 3
[0083] The present invention provides an electronic device, including: a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to perform the forest destruction area identification method of Embodiment 1.
[0084] Example 4
[0085] The present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the forest destruction area identification method of Embodiment 1.
[0086] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the method section.
[0087] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.
Claims
1. A method for identifying deforestation zones, characterized in that, include: Acquire Sentinel-1 GRD data and Sentinel-1 SLC data for the forest area to be identified; The Sentinel-1 GRD data undergoes a first preprocessing step, and the Sentinel-1 SLC data undergoes a second preprocessing step to obtain preprocessed GRD data and preprocessed SLC data. The first preprocessing step includes orbit correction, radiometric correction, thermal noise and edge noise removal, terrain correction, and filtering. The second preprocessing step includes orbit correction, radiometric correction, multi-view, polarization matrix generation, terrain correction, filtering, and polarization decomposition. Based on the preprocessed GRD data and the preprocessed SLC data, the optimal feature parameters for all pixels in the forest region to be identified are calculated. The optimal feature parameters include a first optimal feature parameter, a second optimal feature parameter, and a third optimal feature parameter. The first optimal feature parameter is determined based on the separation of backscattering intensity parameters between the forest destruction area and the environmental background. The backscattering intensity parameters include co-polarized backscattering intensity and cross-polarized backscattering intensity. The second optimal feature parameter is determined based on the separation of vegetation indices between the forest destruction area and the environmental background. The vegetation indices include radar vegetation index, polarimetric radar vegetation index, and radar forest degradation index. The third optimal feature parameter is determined based on the separation of polarization parameters between the forest destruction area and the environmental background. The polarization parameters include the diagonal parameter of the polarization matrix, polarization entropy, polarization anisotropy, and polarization angle. The process of determining the optimal feature parameters specifically includes: Acquire Sentinel 1 GRD sample data and Sentinel 1 SLC sample data in the forest experimental area; The Sentinel 1 GRD sample data is subjected to a first preprocessing, and the Sentinel 1 SLC sample data is subjected to a second preprocessing to obtain preprocessed GRD sample data and preprocessed SLC sample data. Based on the preprocessed GRD sample data, calculate the same polarization backscattering intensity, cross polarization backscattering intensity, radar vegetation index, polarization radar vegetation index, and radar forest degradation index for all pixels in the forest experimental area. Based on the preprocessed SLC sample data, calculate the diagonal parameters of the polarization matrix, polarization entropy, polarization anisotropy and polarization angle of all pixels in the forest experimental area. Calculate the same polarization backscattering intensity, cross polarization backscattering intensity, radar vegetation index, polarization radar vegetation index, radar forest degradation index, polarization matrix diagonal parameter, polarization entropy, polarization anisotropy, and polarization angle of all pixels to determine the separation degree between the forest destruction area and the environmental background in the experimental area. Based on all the said separation degrees, a first optimal feature parameter, a second optimal feature parameter, and a third optimal feature parameter are determined; The calculation of the same-polarization backscattering intensity, cross-polarization backscattering intensity, radar vegetation index, polarimetric radar vegetation index, radar forest degradation index, diagonal parameter of the polarization matrix, polarization entropy, polarization anisotropy, and the separation degree of the polarization angle in the deforestation area and the environmental background for all pixels specifically includes: Using formula Calculate the same-polarization backscattering intensity, the cross-polarization backscattering intensity, the radar vegetation index, the polarization radar vegetation index, the radar forest degradation index, the diagonal parameter of the polarization matrix, the polarization entropy, the polarization anisotropy, and the separation degree of the polarization angle in the forest destruction area and the environmental background; wherein, This represents the average value of a certain feature parameter across all pixels in the deforestation area. It represents the average value of a certain feature parameter of all pixels in the background environment; The standard deviation of a certain feature parameter of all pixels in the deforestation area; The standard deviation of a certain feature parameter of all pixels in the environmental background; Based on the optimal feature parameters of all the pixels, the forest destruction zone of the forest area to be identified is determined using the K-means clustering algorithm.
2. The method for identifying deforestation areas according to claim 1, characterized in that, The number of clusters in the K-means clustering algorithm is set to 2, and the number of iterations is set to 25.
3. The method for identifying deforestation areas according to claim 1, characterized in that, Based on the preprocessed GRD data and the preprocessed SLC data, the optimal feature parameters for all pixels in the forest region to be identified are calculated, specifically including: Based on the preprocessed GRD data, calculate the first optimal feature parameter and the second optimal feature parameter for all pixels in the forest region to be identified. Based on the preprocessed SLC data, the third optimal feature parameters of all pixels in the forest region to be identified are calculated.
4. The method for identifying deforestation areas according to claim 1, characterized in that, Also includes: A three-dimensional feature space model of the forest region to be identified is established based on the optimal feature parameters of all the pixels; the coordinate axes of the three-dimensional feature space model are the first optimal feature parameter, the second optimal feature parameter, and the third optimal feature parameter, respectively.
5. The method for identifying deforestation areas according to claim 1, characterized in that, Based on all the stated separation degrees, a first optimal feature parameter, a second optimal feature parameter, and a third optimal feature parameter are determined, specifically including: The parameter with the highest separation degree among the backscattering intensity parameters is selected as the first optimal characteristic parameter; The parameter with the highest separation degree among the vegetation indices is selected as the second optimal feature parameter. Among the polarization parameters, the parameter with the highest separation degree is selected as the third optimal characteristic parameter.
6. A forest destruction zone identification system, characterized in that, include: The data acquisition module is used to acquire Sentinel-1 GRD data and Sentinel-1 SLC data of the forest area to be identified; The preprocessing module is used to perform a first preprocessing on the Sentinel-1 GRD data and a second preprocessing on the Sentinel-1 SLC data to obtain preprocessed GRD data and preprocessed SLC data. The first preprocessing includes orbit correction, radiometric correction, thermal noise and edge noise removal, terrain correction, and filtering. The second preprocessing includes orbit correction, radiometric correction, multi-view, polarization matrix generation, terrain correction, filtering, and polarization decomposition. The feature calculation module is used to calculate the optimal feature parameters for all pixels in the forest area to be identified based on the preprocessed GRD data and the preprocessed SLC data. The optimal feature parameters include a first optimal feature parameter, a second optimal feature parameter, and a third optimal feature parameter. The first optimal feature parameter is determined based on the separation of backscattering intensity parameters between the forest destruction area and the environmental background. The backscattering intensity parameters include co-polarized backscattering intensity and cross-polarized backscattering intensity. The second optimal feature parameter is determined based on the separation of vegetation indices between the forest destruction area and the environmental background. The vegetation indices include radar vegetation index, polarimetric radar vegetation index, and radar forest degradation index. The third optimal feature parameter is determined based on the separation of polarization parameters between the forest destruction area and the environmental background. The polarization parameters include the diagonal parameter of the polarization matrix, polarization entropy, polarization anisotropy, and polarization angle. The process of determining the optimal feature parameters specifically includes: Acquire Sentinel 1 GRD sample data and Sentinel 1 SLC sample data in the forest experimental area; The Sentinel 1 GRD sample data is subjected to a first preprocessing, and the Sentinel 1 SLC sample data is subjected to a second preprocessing to obtain preprocessed GRD sample data and preprocessed SLC sample data. Based on the preprocessed GRD sample data, calculate the same polarization backscattering intensity, cross polarization backscattering intensity, radar vegetation index, polarization radar vegetation index, and radar forest degradation index for all pixels in the forest experimental area. Based on the preprocessed SLC sample data, calculate the diagonal parameters of the polarization matrix, polarization entropy, polarization anisotropy and polarization angle of all pixels in the forest experimental area. Calculate the same polarization backscattering intensity, cross polarization backscattering intensity, radar vegetation index, polarization radar vegetation index, radar forest degradation index, polarization matrix diagonal parameter, polarization entropy, polarization anisotropy, and polarization angle of all pixels to determine the separation degree between the forest destruction area and the environmental background in the experimental area. Based on all the said separation degrees, a first optimal feature parameter, a second optimal feature parameter, and a third optimal feature parameter are determined; The calculation of the same-polarization backscattering intensity, cross-polarization backscattering intensity, radar vegetation index, polarimetric radar vegetation index, radar forest degradation index, diagonal parameter of the polarization matrix, polarization entropy, polarization anisotropy, and the separation degree of the polarization angle in the deforestation area and the environmental background for all pixels specifically includes: Using formula Calculate the same-polarization backscattering intensity, the cross-polarization backscattering intensity, the radar vegetation index, the polarization radar vegetation index, the radar forest degradation index, the diagonal parameter of the polarization matrix, the polarization entropy, the polarization anisotropy, and the separation degree of the polarization angle in the forest destruction area and the environmental background; wherein, This represents the average value of a certain feature parameter across all pixels in the deforestation area. It represents the average value of a certain feature parameter of all pixels in the background environment; The standard deviation of a certain feature parameter of all pixels in the deforestation area; The standard deviation of a certain feature parameter of all pixels in the environmental background; The identification module is used to determine the forest destruction area of the forest area to be identified by using the K-means clustering algorithm based on the best feature parameters of all the pixels.
7. An electronic device, characterized in that, include: A memory and a processor, the memory being used to store a computer program, the processor running the computer program to cause the electronic device to perform the forest destruction area identification method according to any one of claims 1-5.
8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the forest destruction area identification method according to any one of claims 1-5.