Fine flood inundation area extraction method based on satellite-borne SAR image

Abstract

The invention provides a fine flood inundation area extraction method based on a satellite-borne SAR image, and the method comprises the steps: S1, obtaining an SAR image of a target area, carrying out the superpixel segmentation and threshold segmentation of the SAR image, obtaining a segmentation result, obtaining a background percentage according to the segmentation result, and calculating the feature vector of each image object; S2, constructing an undirected graph by processing the superpixel segmentation result, the background percentage and the feature vector, and obtaining category labels of part of nodes; S3, training a graph convolutional neural network according to the undirected graph and the obtained category labels of the partial nodes, and predicting categories of other nodes through the trained graph convolutional neural network to obtain a submerged water body extraction result; and S4, subtracting the extracted submerged water body distribution from the water body distribution before the flood to obtain a flood submerged area. According to the method, the spatial relationship between the image object features and the image objects can be considered, and the extraction precision of the flood inundation area in the SAR image can be remarkably improved.

Description

technical field [0001] The invention belongs to the technical field of remote sensing image processing and information extraction, and in particular relates to a refined flood submerged area extraction method based on spaceborne SAR images. Background technique [0002] In recent years, the global climate has been abnormal, and floods have occurred frequently. Since last year, my country has experienced the Yangtze River embankment burst, Henan Zhengzhou Xinxiang flood, Inner Mongolia reservoir dam break and Shanxi flood, etc., causing many casualties and incalculable economic losses. The rapid extraction of flood submerged areas is beneficial to emergency relief, disaster loss assessment and post-disaster reconstruction, etc., which has great research significance and is imperative. Remote sensing satellite observation is the main way to observe the earth. Synthetic aperture radar satellites play an important role in emergency disaster reduction because of their significan...

AI Technical Summary

Problems solved by technology

Therefore, this method has the following problems: a) requires a large amount of manually labeled data; b) requires a large amount of computing resources to train the network; c) needs to be deployed on a high-performance platform

At present, the methods for extracting inundated areas are mainly the above three methods, and these methods are mainly aimed at ordinary water body areas, and cannot be fully applied to flooded areas

Therefore, there is a lack of fine extraction methods for flooded areas, especially those using high-resolution SAR images

Images