SAR image glacier recognition method based on texture feature assistance

Abstract

The invention relates to an SAR (Synthetic Aperture Radar) image glacier recognition method based on texture feature assistance, which is used for recognizing a glacier by utilizing an interference measurement method of Sentinel-1A data and combining texture information. The method comprises the following steps: utilizing Sentinel-1A data to carry out interference processing to obtain a coherencecoefficient graph; extracting seven textural features of mean value, variance, homogeneity, contrast, dissimilarity, entropy and correlation of the coherence coefficient graph; and comparing glacier extraction effects among different texture feature combinations. And selecting four textural features of mean value, variance, homogeneity and dissimilarity under the VV polarization mode with the highest classification precision, and training by utilizing an SVM classifier to obtain classes of glaciers, rocks, accumulated snow and the like. And converting the classified data into a shape format, and segmenting the data through ridge lines to obtain a glacier boundary data set. Compared with a traditional threshold segmentation method based on a coherence coefficient graph, texture features areintroduced, and glacier recognition is more accurate.

Description

technical field [0001] The invention relates to a SAR image glacier recognition method based on texture features, and belongs to the technical field of remote sensing geoscience applications. Background technique [0002] Glacier resources are important freshwater resources on the earth. As an important indicator of global climate change, glacier identification using remote sensing methods is of great significance for monitoring glacier area and its reserves, and studying global climate change. However, since there is a layer of debris mixture on the surface of moraine-covered glaciers, which have similar spectral information to the rocks around the glacier in optical images, optical remote sensing identification methods such as band ratio method and normalized snow cover index can only identify relatively clean The pure glacier in the area, and the optical remote sensing is greatly affected by the weather conditions, so it is difficult to observe the same area at the same t...

AI Technical Summary

Problems solved by technology

However, since there is a layer of debris mixture on the surface of moraine-covered glaciers, which have similar spectral information to the rocks around the glacier in optical images, optical remote sensing identification methods such as band ratio method and normalized snow cover index can only identify relatively clean The pure glacier in the area, and the optical remote sensing is greatly affected by the weather conditions, so it is difficult to observe the same area at the same time

Images