CN-121982573-A - Remote sensing image monitoring method and system for soil carbon loss in hydraulic erosion area

CN121982573ACN 121982573 ACN121982573 ACN 121982573ACN-121982573-A

Abstract

The invention discloses a method and a system for monitoring a soil carbon loss remote sensing image in a hydraulic erosion area, which belong to the field of remote sensing image processing, wherein the method comprises the steps of extracting spectral characteristics by utilizing hyperspectral images, and inverting the organic carbon content of soil by adopting a PLSR and SVM algorithm construction model; the method comprises the steps of superposing a grading graph of carbon content and erosion strength, calculating a carbon loss rate, establishing a coupling relation, constructing a source-sink index by combining a topographic parameter, dividing an erosion carbon output area, a migration transition area and a deposition carbon enrichment area, comparing the carbon content change before and after a snow melting period and a rainy season, and calculating seasonal carbon loss.

Inventors

LIU ZIJIN
SI JIANHUA
XIAO SHENGCHUN
JIA BING
Zhou Dongmeng
ZHU XINGLIN
BAI XUE
WANG BOYANG

Assignees

中国科学院西北生态环境资源研究院

Dates

Publication Date: 20260505
Application Date: 20260131

Claims (10)

1. The remote sensing image monitoring method for soil carbon loss in the hydraulic erosion area is characterized by comprising the following steps of: s1, spectrum inversion is carried out, namely a hyperspectral remote sensing image of a research area is obtained, the hyperspectral remote sensing image is preprocessed to obtain spectral reflectivity data, spectral characteristics of visible light-near infrared wave bands are extracted, exposed soil pixels are identified, a spectrum inversion model is built by combining ground sampling point soil organic carbon content actual measurement data and adopting partial least square regression and a support vector machine algorithm, and the soil organic carbon content of the surface layer of the exposed soil is estimated in a planar mode to obtain a soil carbon content inversion result; S2, a coupling analysis step, namely acquiring an erosion strength grading diagram of a research area, carrying out space superposition analysis on soil carbon content inversion results and the erosion strength grading diagram, calculating carbon content statistical characteristic values and carbon content difference values of different erosion grade areas, calculating a carbon loss rate based on the erosion strength and the carbon content difference values, and establishing an erosion-carbon loss coupling relation; Step S3, a source and sink identification step, namely acquiring digital elevation model data of a research area, extracting terrain gradient, terrain humidity index and sink accumulation amount terrain parameters, constructing a source and sink pattern discrimination index by combining erosion deposition characteristics, dividing the research area into an erosion carbon output area, a migration transition area and a deposition carbon enrichment area, and identifying a space redistribution pattern of organic carbon in a river basin; And S4, a time sequence monitoring step, namely acquiring multi-temporal hyperspectral remote sensing images before and after a snow melting period and before and after a rain season, selecting four key time-phase images, respectively executing the steps S1 to S3, comparing the spatial distribution changes of the carbon content of soil in different time phases, respectively accounting the loss amount of carbon in snow melting erosion and the loss amount of carbon in rainfall erosion, and outputting a carbon loss space-time dynamic monitoring result.
2. The method for monitoring soil carbon loss remote sensing image in a hydraulically eroded area according to claim 1, wherein in step S1, the spectral features are extracted in the wavelength bands of 480nm to 520nm, 580nm to 680nm, 1350nm to 1450nm, 1850nm to 1950nm and 2150nm to 2250 nm.
3. The method for monitoring soil carbon loss remote sensing images in a hydraulically eroded area according to claim 1, wherein in step S1, bare soil pixels are identified using a threshold value with normalized vegetation index NDVI less than 0.2.
4. The method for monitoring soil carbon loss remote sensing images in a hydraulic erosion area according to claim 1, wherein in step S2, the difference value of carbon content is calculated based on the micro erosion area, and the average difference value of carbon content between each grade of erosion area and the micro erosion area is calculated.
5. The method for monitoring a remote sensing image of soil carbon loss in a hydraulic erosion area according to claim 1, wherein in step S2, the calculation of the carbon loss rate is obtained by combining erosion-carbon loss coupling coefficient, soil erosion modulus, soil volume weight, effective erosion depth and soil organic carbon content calculation.
6. The method for monitoring a remote sensing image of soil carbon loss in a hydraulic erosion area according to claim 1, wherein in step S3, the source pattern discrimination index comprehensively considers a difference value between a topography gradient and a critical gradient, a normalized value of a topography humidity index and a normalized deviation value of a soil organic carbon content.
7. The method for monitoring a remote sensing image of soil carbon loss in a hydraulic erosion area according to claim 6, wherein in the step S3, the method is divided into erosion carbon output areas when the source-sink pattern discrimination index is greater than 0.3, migration transition areas when the source-sink pattern discrimination index is between-0.3 and 0.3, and sediment carbon enrichment areas when the source-sink pattern discrimination index is less than-0.3.
8. The method for monitoring the remote sensing image of the soil carbon loss in the hydraulic erosion area according to claim 1, wherein in the step S4, the images of four key time phases are selected from the group consisting of a snow melting early stage, a snow melting later stage, a rainy season early stage and a rainy season later stage.
9. The method for monitoring soil carbon loss remote sensing images in a hydraulic erosion area according to claim 1, wherein in step S4, the calculation of the snow-melting erosion carbon loss and the rainfall erosion carbon loss is performed by using a snow-melting erosion carbon loss efficiency coefficient and a rainfall erosion carbon loss efficiency coefficient, respectively, wherein the value of the snow-melting erosion carbon loss efficiency coefficient ranges from 0.6 to 0.9, and the value of the rainfall erosion carbon loss efficiency coefficient ranges from 0.7 to 0.95.
10. A remote sensing image monitoring system for soil carbon loss in a hydraulic erosion area, for implementing the remote sensing image monitoring method for soil carbon loss in a hydraulic erosion area according to any one of claims 1 to 9, comprising: the data acquisition module is used for acquiring hyperspectral remote sensing images, erosion intensity grading diagrams, digital elevation model data and ground sampling point soil organic carbon content actual measurement data of a research area; the spectrum inversion module is used for preprocessing the hyperspectral remote sensing image to obtain spectral reflectivity data, extracting spectral characteristics of visible light-near infrared bands, constructing a spectrum inversion model by adopting partial least square regression and a support vector machine algorithm, and obtaining a soil carbon content inversion result; the coupling analysis module is used for carrying out space superposition analysis on soil carbon content inversion results and the erosion strength grading graph, calculating carbon content statistical characteristic values and carbon content difference values of different erosion grade areas, and calculating the carbon loss rate based on the erosion strength and the carbon content difference values; The source-sink identification module is used for extracting topographic parameters and combining erosion deposition characteristics to construct a source-sink pattern discrimination index, and dividing a research area into an erosion carbon output area, an migration transition area and a deposition carbon enrichment area; the time sequence monitoring module is used for acquiring multi-time-phase hyperspectral remote sensing images, and respectively accounting the snow melting erosion carbon loss and the rainfall erosion carbon loss by comparing the spatial distribution changes of the carbon content of the soil in different time phases.

Description

Remote sensing image monitoring method and system for soil carbon loss in hydraulic erosion area Technical Field The invention relates to the technical field of remote sensing image processing and water and soil conservation monitoring, in particular to a method and a system for monitoring a remote sensing image of soil carbon loss in a hydraulic erosion area. Background Hydraulic erosion is one of the important driving factors for the loss of organic carbon in soil, especially in semiarid alpine mountain areas, and the soil erosion process is accompanied by significant transverse migration of organic carbon due to the dual effects of seasonal freeze thawing cycles and concentrated rainfall. It is counted that the loss of organic carbon in soil caused by hydraulic erosion is as high as billions tons each year worldwide, and a considerable proportion of the loss occurs in ecological fragile areas such as alpine mountainous areas. Soil organic carbon is taken as a largest carbon warehouse component of the land ecological system, and loss of the soil organic carbon not only leads to soil fertility reduction and ecological function degradation, but also can release carbon dioxide through mineralization to exacerbate greenhouse effect. Therefore, the space-time dynamics of soil carbon loss in the hydraulic erosion area is accurately monitored, and the method has important scientific value and practical significance for regional carbon circulation research and water and soil conservation carbon sink benefit evaluation. The migration and transformation of soil organic carbon in the erosion process is a complex physicochemical process, and involves multiple links such as stripping, transportation, deposition, mineralization and the like of carbon. In the erosion generation area, the soil with the surface layer rich in organic carbon is stripped by runoff and moves downwards, so that the in-situ carbon content is reduced, and in the deposition area, sediment deposition and enrichment of the organic carbon are carried, so that the carbon content of the area is relatively increased. This erosion-driven carbon lateral redistribution process profoundly affects the balance of carbon balance on the watershed scale. Particularly in alpine mountain areas, the snow melting period in spring and the rain season in summer are two main active periods of hydraulic erosion, and snow melting water and rain runoff drive carbon loss processes with different intensities and space patterns respectively. The remote sensing technology has become an important means for soil erosion and carbon cycle monitoring because of the advantages of large-scale, rapid and nondestructive acquisition of surface information. The existing soil erosion remote sensing monitoring technology is mainly focused on the extraction and analysis of erosion morphological characteristics. For example, chinese patent application publication No. CN116934753a discloses a soil and water conservation monitoring method based on a remote sensing image, which extracts an edge line of an erosion trench by performing edge detection on a high-resolution remote sensing image, constructs a gully span ratio according to the direction and width change of the erosion trench, constructs a single-sided gully sag index by combining the corresponding relation between the depth of the erosion trench and an image gray value, and calculates the soil erosion saliency to evaluate the soil erosion degree. The method improves the image segmentation precision of the soil erosion area to a certain extent, and provides technical support for water and soil conservation effect evaluation. In the aspect of soil organic carbon remote sensing inversion, the development of the hyperspectral remote sensing technology provides a new approach for planar estimation of the carbon content of soil. The soil organic matter has characteristic absorption characteristics in visible light and near infrared bands, and by utilizing the spectrum sensitivity response characteristic, the quick estimation of the carbon content of the soil can be realized by establishing a spectrum inversion model. Machine learning algorithms such as partial least squares regression and support vector machines are widely applied to spectral modeling of soil carbon content, and good prediction effects are obtained. However, existing research focuses on inversion of carbon content in soil in farmlands or plain areas, research on hydraulic erosion areas, particularly alpine mountains, is relatively scarce, and there are few systematic methods that combine inversion of carbon content with analysis of erosion processes. However, the above-described prior art still has the following disadvantages. Firstly, the existing method mainly focuses on physical morphological characteristics of soil erosion, such as geometric parameters of erosion furrows, but fails to establish quantitative association between an erosion process and soil organic carbo