Soil organic carbon high-precision inversion method based on habitat plaque earth surface unit division
Through the method based on habitat patch surface unit division, combined with multi-source remote sensing data and machine learning model, the problem of low accuracy of soil organic carbon remote sensing estimation in complex terrain areas is solved, and high-precision remote sensing of soil organic carbon content in complex terrain areas is achieved in the southwest Estimate.
Patent Information
- Application Number
- CN202410931264.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-12
- Publication Date
- 2024-11-08
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
In complex terrain areas, it is difficult for the prior art to achieve high-precision remote sensing estimation of soil organic carbon, and there is a lack of fine estimation methods for remote sensing of soil organic carbon for complex terrain areas.
Using a method based on habitat patch surface unit division, the multi-source environmental remote sensing data is obtained for preprocessing, the spectral index is calculated, and the soil organic carbon content prediction model is established in different sub-regions using clustering analysis and machine learning model, combining support vectors For predictions, random forests, extreme gradient lift decision trees and lightweight gradient lift models.
The remote sensing fine estimation of soil organic carbon content in complex terrain areas in the southwest has been achieved, the simulation accuracy is improved, and the problem of low simulation accuracy of a single model is compensated.
Smart Images

Figure CN118916768A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of ecological environment assessment, and specifically relates to a high-precision inversion method for soil organic carbon based on habitat patch surface unit division. Background Art
[0002] Rapidly acquiring high-resolution and large-scale spatial information of soil organic carbon, accurately estimating its spatial distribution, and understanding its spatial variation are of great significance for mitigating global climate change, developing precision agriculture, evaluating soil carbon sequestration potential, and improving environmental quality. However, high-precision digital soil mapping in complex terrain areas (with high spatial heterogeneity) is extremely challenging. In terms of soil organic carbon variability analysis, soil sampling and field surveys have traditionally been the main data acquisition methods. With the increasing complexity of soil mapping requirements and the emergence of high-temporal and spatial resolution remote sensing data, satellite remote sensing data has been increasingly used in soil organic carbon research. Using remote sensing data to obtain information such as climate factors, terrain factors, soil types, optical environment, and microwaves, as well as human activity information such as land use changes and population density, combined with geostatistical and spatial analysis methods, the spatial distribution pattern of soil organic carbon in complex terrain areas can be obtained. However, most current methods are based on modeling and estimating soil organic carbon in the entire study area. There are still some deficiencies in remote sensing estimation of soil organic carbon content in complex terrain areas, and there is a lack of remote sensing fine estimation methods for soil organic carbon in complex terrain areas. Summary of the invention
[0003] In order to solve the above problems, the present invention provides a high-precision inversion method of soil organic carbon based on habitat patch surface unit division. The specific technical scheme is as follows:
[0004] The high-precision inversion method of soil organic carbon based on habitat patch surface unit division includes the following steps:
[0005] Step S1, firstly obtain multi-source environment-remote sensing data of the study area and pre-process the image data;
[0006] Firstly, the polarization data of Sentinel-1 and Landsat8 image data of the study area were collected, and the measured data of SOC, annual average temperature, annual precipitation, land surface temperature LST, population density, land use, DEM and its extracted four related terrain indicators were collected.
[0007] The preprocessing includes radiation calibration, atmospheric correction, image stitching and cropping of Landsat8 image data; meanwhile, the SOC measured data includes SOC collection data and SOC field sampling data.
[0008] The terrain indicators including terrain relief TU, slope, aspect and terrain wetness index TWI were extracted from the 30 m digital elevation model DEM.
[0009] Step S2, calculating the spectral index of the research area according to the preprocessed Landsat8 data;
[0010] The spectral indices of the study area were obtained through band calculation, including brightness index BI, second brightness index BI2, color index CI, clay index CI1, ratio vegetation index RVI, green-red vegetation index GRVI, green normalized difference vegetation index GNDVI, surface water index LSWI, secondary corrected soil corrected vegetation index MSAVI2, normalized water index NDWI, modified normalized water index MNDWI, water stress index MSI, surface water index LSWI, redness index RI, soil-adjusted total vegetation index SATVI, soil-corrected vegetation index SAVI, transformed vegetation index TVI, leaf area index LAI, normalized vegetation index NDVI, green normalized vegetation index GNDVI and enhanced vegetation index EVI, totaling 21 indices.
[0011] In step S3, we first proposed the concept of habitat patches and the division method of habitat patches as surface units. The specific concept is: "a specific geographical space of an ecological environment formed under the influence of similar natural and human factors, and the internal patches have relative uniformity of environmental factors, including the state of environmental factors and the characteristics of temporal changes". The specific division method includes but is not limited to spatial clustering analysis, environmental stratification, land use type and different surface unit division combination methods. In the present invention, the preferred spatial clustering analysis method is: the study area is divided into four different habitat patch types using the clustering algorithm around the central partition PAM; after this division, the effects of optimizing the auxiliary variable combination of each area using filters, wrappers and embedded feature selection methods are compared;
[0012] Step S4, based on the regional division results after cluster analysis and the factor screening results of the optimal feature selection method, a variety of machine learning models are used to establish soil organic carbon content prediction models in different sub-regions and the entire region; based on the comparison of prediction accuracy of different models, a spatial distribution map of soil organic carbon content in the complex terrain area of the southwest is generated.
[0013] Furthermore, in step S4, the machine learning models are support vector machine SVM, random forest RF, extreme gradient boosting decision tree XGBoost and lightweight gradient boosting machine LightGBM.
[0014] The accuracy of the simulated output results of the machine learning model was measured by R 2, RMSE, MAE and RPD are evaluated. The specific evaluation formula is as follows:
[0015]
[0016]
[0017]
[0018]
[0019] Where P i and O i are the predicted and measured soil organic carbon contents, n is the number of sample points, and are the average values of predicted and measured soil organic carbon contents, respectively, and SD is the standard deviation of the measured soil organic carbon contents.
[0020] The beneficial effects of the present invention are as follows: based on full consideration of environmental variables, remote sensing optical indices and microwave remote sensing data, the present invention introduces a cluster analysis method to reasonably divide the study area in view of the strong surface heterogeneity of the study area, and constructs a soil organic carbon remote sensing estimation model suitable for the complex terrain areas in the southwest, thereby realizing the remote sensing fine estimation of soil organic carbon content in the complex terrain areas in the southwest, and making up for the problem that the use of a single model to simulate SOC in existing complex terrain areas may lead to low simulation accuracy. BRIEF DESCRIPTION OF THE DRAWINGS
[0021] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following is a brief introduction to the drawings required for the specific embodiments or the prior art description. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the parts are not necessarily drawn according to the actual scale.
[0022] Figure 1 It is a schematic diagram of the technical process of the present invention;
[0023] Figure 2 This is a regional schematic diagram of a complex terrain area selected for research in a specific embodiment of the present invention. DETAILED DESCRIPTION
[0024] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0025] It should be understood that when used in this specification and the appended claims, the terms "include" and "comprises" indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.
[0026] It should also be understood that the terms used in the present specification are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.
[0027] It should be further understood that the term "and / or" used in the present description and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.
[0028] like Figure 1 As shown, a specific embodiment of the present invention provides a method for high-precision estimation of soil organic carbon in complex terrain areas in the southwest, comprising the following steps:
[0029] Step S1, this embodiment selects Figure 2 The southwestern region shown in the figure is used as the complex terrain research area of the present invention. First, the Landsat8 image data of the southwestern region is collected and the image data is preprocessed; the preprocessing includes radiometric calibration, atmospheric correction, image stitching and cropping of the Landsat8 image data. The VV and VH polarization data of Sentinel-1, annual average temperature, annual precipitation, surface temperature LST, population density, land use, DEM and its extracted terrain undulation TU, slope, slope aspect and terrain wetness index TWI 4 related terrain indicators are collected.
[0030] Step S2, according to the preprocessed Landsat8 data, the spectral index of the southwestern region is calculated, including the brightness index BI, the second brightness index BI2, the color index CI, the clay index CI1, the ratio vegetation index RVI, the green-red vegetation index GRVI, the green normalized difference vegetation index GNDVI, the surface water index LSWI, the secondary corrected soil corrected vegetation index MSAVI2, the normalized water index NDWI, the modified normalized water index MNDWI, the water stress index MSI, the surface water index LSWI, the redness index RI, the soil-adjusted total vegetation index SATVI, the soil-corrected vegetation index SAVI, the transformed vegetation index TVI, the leaf area index LAI, the normalized vegetation index NDVI, the green normalized vegetation index GNDVI and the enhanced vegetation index EVI, a total of 21 indices, as shown in Table 1.
[0031] Among them, B2, B3, B4, B5, B6 and B7 represent the reflectivity of the blue band (450-515nm), green band (525-600nm), red band (630-680nm), near-infrared band (845-885nm), short-wave infrared band (1560-1660nm) and short-wave infrared band 2 (2100-2300nm) corresponding to the Landsat8 image, respectively.
[0032] In step S3, the study area is divided into four different habitat patches using the Partitioning Around Centers (PAM) algorithm. After this division, the effects of optimizing the auxiliary variable combination for each region using filters, wrappers, and embedded feature selection methods are compared.
[0033] Step S4, based on the regional division results after cluster analysis and the factor screening results of the optimal feature selection method, a machine learning model and a neural network model are used to establish soil organic carbon content prediction models in different sub-regions and the entire region respectively; based on the comparison of the prediction accuracy of different models, a spatial distribution map of soil organic carbon content in the complex terrain area of the southwest is generated.
[0034] Table 1 Definition of spectral index
[0035]
[0036] Those of ordinary skill in the art will appreciate that the units of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition of each example has been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.
[0037] In the embodiments provided in the present application, it should be understood that the division of habitat patch units is only a logical functional division, and there may be other division methods in actual implementation, for example, multiple units can be combined into one unit, one unit can be split into multiple units, different unit combinations or some features can be ignored, etc.
[0038] Finally, it should be noted that the above 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 aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents. These modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be included in the scope of the claims and specification of the present invention.
Claims
1. A high-precision inversion method for soil organic carbon based on habitat patch surface unit division, characterized in that: The following steps are involved: Step S1, firstly obtain multi-source environment-remote sensing data of the study area and pre-process the image data; First, we collected the polarization data of Sentinel-1 and Landsat8 image data of the study area, and collected the measured data of SOC, annual average temperature, annual precipitation, land surface temperature LST, population density, land use, DEM and its extracted related terrain indicators; Step S2, calculating the spectral index of the research area according to the preprocessed Landsat8 data; Step S3, using the habitat patch surface unit division method; the habitat patch concept is: an ecological environment-specific geographical space formed under the action of similar natural and human factors, and the patch has relative uniformity of environmental factors, including the state of environmental factors and time change characteristics; the study area is divided into four different habitat patch types; after the division, the effect of optimizing the auxiliary variable combination of each area by using filters, wrappers and embedded feature selection methods is compared; Step S4, based on the regional division results after cluster analysis and the factor screening results of the optimal feature selection method, a variety of machine learning models are used to establish soil organic carbon content prediction models in different sub-regions and the entire region; based on the comparison of prediction accuracy of different models, a spatial distribution map of soil organic carbon content in the southwest region is generated.
2. The high-precision inversion method for soil organic carbon based on habitat patch surface unit division according to claim 1 is characterized in that: In step S1, the preprocessing includes performing radiation calibration, atmospheric correction, image stitching and cropping on the Landsat8 image data; meanwhile, the SOC measured data includes SOC collection data and SOC field sampling data.
3. The high-precision inversion method for soil organic carbon based on habitat patch surface unit division according to claim 1 is characterized in that: In step S1, terrain indicators are extracted from the 30 m digital elevation model DEM, including terrain relief TU, slope, aspect and terrain wetness index TWI.
4. The high-precision inversion method for soil organic carbon based on habitat patch surface unit division according to claim 1 is characterized in that: The spectral index of the study area in step S2 is obtained through band operation, including brightness index BI, second brightness index BI2, color index CI, clay index CI1, ratio vegetation index RVI, green-red vegetation index GRVI, green normalized difference vegetation index GNDVI, surface water index LSWI, secondary corrected soil corrected vegetation index MSAVI2, normalized water index NDWI, corrected normalized water index MNDWI, water stress index MSI, surface water index LSWI, redness index RI, soil adjusted total vegetation index SATVI, soil corrected vegetation index SAVI, transformed vegetation index TVI, leaf area index LAI, normalized vegetation index NDVI, green normalized vegetation index GNDVI and enhanced vegetation index EVI, totaling 21 indices.
5. The high-precision inversion method for soil organic carbon based on habitat patch surface unit division according to claim 1 is characterized in that: The specific division method in step S2 adopts spatial cluster analysis, environmental stratification method, land use type or different surface unit division combination method.
6. The high-precision inversion method for soil organic carbon based on habitat patch surface unit division according to claim 1 is characterized in that: In step S4, the machine learning models are support vector machine SVM, random forest RF, extreme gradient boosting decision tree XGBoost and lightweight gradient boosting machine LightGBM.
7. The high-precision inversion method for soil organic carbon based on habitat patch surface unit division according to claim 6 is characterized in that: In step S4, the accuracy of the simulated output results of the machine learning model is measured by R 2 , RMSE, MAE and RPD are evaluated. The specific evaluation formula is as follows: Where P i and O i are the predicted and measured soil organic carbon contents, n is the number of sample points, and are the average values of predicted and measured soil organic carbon contents, respectively, and SD is the standard deviation of the measured soil organic carbon contents.
Citation Information
Patent Citations
Soil organic carbon estimation method and device based on multi-source remote sensing data
CN114740180A
High-precision estimation method for soil organic carbon in karst trough valley area
CN117746001A
Cited By
Nonlinear evolution path and characteristic evaluation method for organic carbon in regional soil and application of nonlinear evolution path and characteristic evaluation method
CN120219856A
Identification method of soil organic carbon multi-level management unit and related device
CN120298904A
Method and system for predicting spatial distribution of organic carbon in soil based on multispectral satellite remote sensing data
CN121685982A
A method and system for predicting the spatial distribution of soil organic carbon based on multispectral satellite remote sensing data
CN121685982B
