Earthquake Damage Remote Sensing Image Segmentation Method and System Based on Dynamic Chain Graph Model

Abstract

The invention discloses a method and system for segmenting earthquake damage remote sensing images based on a dynamic chain graph model, wherein the method includes initial segmentation of multi-spectral earthquake damage remote sensing images to obtain initial segmentation regions of the multi-spectrum earthquake damage remote sensing images; Calculate the degree of heterogeneity between all initial segmentation regions; construct a chain graph model according to the degree of heterogeneity between the segmentation regions and the adjacent relationship between the segmentation regions; the chain graph model includes interlinked region adjacency graphs and nearest neighbors Graph; use the edge length in the nearest neighbor graph as the primary key to construct the priority queue based on the red-black tree, and dynamically merge the priority queue based on the red-black tree according to the rule of the lowest heterogeneity and the first priority, and finally get the corresponding earthquake-damaged ground objects. Segmentation results of matched multispectral earthquake damage remote sensing images. The invention effectively avoids erroneous segmentation in the complex earthquake damage remote sensing image segmentation, improves the segmentation accuracy rate, and makes the segmentation result more match with the earthquake damage ground objects.

Description

technical field [0001] The invention belongs to the technical field of remote sensing image processing, and in particular relates to a method for segmenting earthquake damage remote sensing images based on a dynamic chain graph model and a system thereof. Background technique [0002] Earthquake disasters occur frequently and bring great losses to human beings. In the face of severe earthquake disasters, how to quickly and accurately extract the disaster information is of great significance to provide reliable information support for disaster assessment, earthquake relief, post-disaster reconstruction, etc. [0003] The development of high-resolution earth observation technology provides data guarantee for earthquake disaster monitoring. At present, disaster interpretation based on high-score data is mainly based on visual judgment and manual interpretation, and the degree of automation is not high, resulting in low efficiency and strong subjectivity. Image segmentation is...

AI Technical Summary

Problems solved by technology

[0004] Although the image segmentation method based on region merging has significant advantages over other methods, there are still some shortcomings in the face of remote sensing earthquake damage images with large data volume and high complexity: (1) the algorithm is inefficient; (2) The merging criterion lacks robustness; (3) the order of regional merging is disordered, prone to wrong segmentation, and it is difficult to ensure the minimum overall segmentation error; (4) the boundary positioning error is large, resulting in a mismatch between the segmentation result and the earthquake damage

Images