CN-121999377-A - Wetland degradation monitoring method based on long-time sequence remote sensing data

CN121999377ACN 121999377 ACN121999377 ACN 121999377ACN-121999377-A

Abstract

The patent name is 'a wetland degradation monitoring method based on long-time sequence remote sensing data', and relates to a wetland degradation monitoring and evaluating technology. The conventional wetland degradation monitoring method mostly depends on a single index or a single time point, and cannot comprehensively and stably evaluate the degradation trend of the wetland. According to the invention, a comprehensive wetland degradation evaluation system is constructed by combining long-time sequence remote sensing data with multidimensional evaluation indexes such as wetland water, soil and vegetation, and weights of the indexes are determined based on an Analytic Hierarchy Process (AHP) to obtain a Wetland Degradation Index (WDI). And identifying the time trend and the significance of the wetland degradation by using a regression analysis method. Finally, through spatial analysis and data derivation, the precise identification and the visualization of spatial distribution of the wetland degradation area are realized, and scientific basis is provided for the protection and recovery of the wetland. The invention has higher accuracy, expandability and operability, and provides powerful data support for global wetland protection.

Inventors

YAO YUNLONG
ZHAO CHENXI
WANG WENJI
GUAN YINA
Hao boya
LI XI
ZHANG XINMIAO
YUAN LIN

Assignees

东北林业大学

Dates

Publication Date: 20260508
Application Date: 20251231

Claims (9)

1. A wetland degradation monitoring method based on long-time sequence remote sensing data comprises the following steps: (1) Acquiring remote sensing image data and preprocessing, including wetland range mask, cloud mask processing and image standardization; (2) Calculating key indexes such as humidity index (WET), vegetation condition index (VC), aboveground biomass (AGB), wetland water index (MNDWI), normalized humidity index (NDMI) and the like through long-time sequence remote sensing data, and carrying out normalization treatment on the key indexes; (3) Determining the weight of each index by adopting an Analytic Hierarchy Process (AHP) based on the indexes, and calculating a Wetland Degradation Index (WDI); (4) Calculating the time trend of wetland degradation through regression analysis to obtain the degradation trend slope of each pixel; (5) Performing significance test and identifying a significant degradation trend region; (6) And carrying out spatial analysis and visualization processing on the obvious degradation region to generate a spatial distribution map of the wetland degradation region.
2. The method of claim 1, wherein the wetland range mask performs extraction of wetland class pixels (encoded as 90) by using WorldCover data sets provided by the European Space Agency (ESA) and applies them as mask layers to the remote sensing image data.
3. The method of claim 1, wherein the image normalization ensures consistency of surface reflectance values for different satellite image data by applying different gain coefficients to image data for different Landsat satellites (Landsat 4/5/7 and Landsat 8/9).
4. The method of claim 1, wherein the humidity index (WET) is calculated from band data of Landsat, calculated using a specific coefficient, and normalized by robustNormalize.
5. The method of claim 1, wherein the Wetland Degradation Index (WDI) is calculated by the formula: wherein VC is a vegetation condition index, MNDWI is a wetland water index, WET is a humidity index, and all input variables are normalized values.
6. The method of claim 1, wherein the regression analysis uses an lm function to perform regression analysis on the WDI time series for each pixel point, calculating the slope (representing the change in trend) and p-value (for significance test) of the regression model.
7. The method of claim 1, wherein the saliency test judges the saliency of the wetland degradation trend by a p value (p < 0.05) and identifies a region of the saliency trend by a saliency mask.
8. The method of claim 1, wherein the spatial analysis includes ensuring that analysis results are limited to a study area range by spatial clipping techniques and using a visualization layer to reveal spatial distribution and trend graphs of wetland degradation areas.
9. The method of claim 1, wherein the data export exports WDI, trend graph layer, and saliency results as GeoTIFF files through export.

Description

Wetland degradation monitoring method based on long-time sequence remote sensing data Technical Field The invention relates to the field of ecological environment protection, in particular to a wetland degradation monitoring and evaluating technology, and especially relates to a method for identifying and monitoring a wetland degradation area based on long-time sequence remote sensing data. Background The wetland ecosystem is an important component of the global ecosystem, and the wetland degradation is represented by the deterioration of hydrologic conditions, the reduction of humidity and water content, the reduction of vegetation coverage and vitality and the like, and the degradation of the wetland ecosystem can lead to the loss of biodiversity, the instability of water resources and the reduction of ecological functions. The existing method depends on a single year or a single index, is difficult to stably and objectively describe the degradation trend and quantify the significance of the degradation trend in the multiple years of age, and meanwhile, quality problems such as cloud/shadow/snow and the like and multi-satellite radiometric calibration differences easily cause incomparable time sequence data. Therefore, the existing method has certain limitations in terms of accuracy, real-time performance and comprehensive evaluation capability. Disclosure of Invention The invention aims to provide a wetland degradation monitoring method based on long-time sequence remote sensing data. According to the method, the multi-source remote sensing data and the wetland degradation index are utilized, and the comprehensive evaluation and the spatial analysis of the wetland degradation are carried out by combining the multidimensional evaluation indexes such as the wetland water, the soil and the vegetation, so that the accurate detection of the wetland degradation area is realized. Detailed Description The invention provides a wetland degradation monitoring method based on long-time sequence remote sensing data. Through multi-dimensional analysis of long-time sequence remote sensing image data, the method realizes quantitative evaluation of the wetland degradation area, and comprises the following specific implementation steps: 1. remote sensing data acquisition and preprocessing The method defines a research area through a vector file, and ensures that the selected area covers all wetland samples. In order to improve the accuracy of data processing and ensure the accuracy of analysis results, a wetland-range masking operation is further performed on the research area. For this purpose, a WorldCover dataset provided by the European Space Agency (ESA) is used, which provides global land cover type information, wherein the wetland categories are explicitly marked. In performing the masking operation, first, the wetland category pixels (encoded as 90) are extracted from the WorldCover dataset and applied as a mask layer to the remote sensing image data. This masking step ensures that only the image data of the wetland region is used for subsequent analysis, excluding disturbances of non-wetland regions, so that the calculation of WDI is limited to the wetland region only. This approach significantly enhances the pertinence and accuracy of the analysis. In the image preprocessing stage, in order to ensure the data quality, firstly, cloud mask processing is carried out on the images in the image preprocessing process, and the masks of cloud, shadow and snow are extracted through maskClouds functions, so that the bad images are ensured not to influence the subsequent analysis. The cloud, shadow and snow masks are extracted based on QA bit data of images, so that interference of the atmospheric factors is effectively avoided. Then, an image normalization process is performed. Because image data of different Landsat satellites (including Landsat 4/5/7 and Landsat 8/9) are used, the image data are standardized through SCALEIMAGE functions, and the consistency of the surface reflectivity values of the image data of different satellites is ensured. According to the characteristics of different satellites, different gain coefficients are respectively applied to Landsat 4/5/7 and Landsat 8/9 for processing so as to convert the same into a consistent surface reflectivity value, and a foundation is laid for subsequent analysis. The method uses Landsat remote sensing image data in 1984 so far, and carries out median synthesis based on annual growth season images, so as to construct a long-time sequence data set. These datasets will provide the necessary spatial and temporal dimension support for the calculation of subsequent WDI time series, ensuring timeliness and reliability of the study results. All data acquisition and processing are performed through a Google EARTH ENGINE (GEE) platform, and the powerful remote sensing data processing and analyzing capability of the data acquisition and processing system is fully utilized.