Multi-temporal remote sensing image change detection method based on joint dictionary learning

Abstract

The multi-temporal remote sensing image change detection method based on joint dictionary learning includes: 1) extracting a large number of unchanged samples from the multi-temporal remote sensing image, and performing joint dictionary learning on the samples to obtain the basis of the unchanged samples; 2) combining step 1) The remaining multi-temporal samples that are not selected are used as test set samples; the test set samples are sparsely reconstructed with the base of the unchanged sample; the difference between the test set samples and the reconstructed test set samples is obtained; 3) from the multi-temporal remote sensing image Select a small number of changed samples; use the basis of unchanged samples to sparsely reconstruct the changed samples of different time phases; use the difference between the reconstructed images of different time phase changed samples, and obtain the change threshold of the changed samples through pooling operation; 4 ) to identify the change area of ​​the multi-temporal remote sensing image by combining the difference image and the change threshold of the change sample, and count the detection rate. The invention can greatly reduce the use of marked samples, does not need to manually select the change threshold, and can improve the detection rate of remote sensing image changes.

Description

technical field [0001] The invention belongs to the technical field of information processing, and relates to a multi-temporal multi-spectral image change detection method, in particular to a multi-temporal remote sensing image change detection method based on joint dictionary learning. Background technique [0002] Since the 20th century, the development of information technology and space technology has profoundly changed the way humans observe the earth. "If you want to see a thousand miles, go to a higher level", since the launch of the first artificial satellite, human beings have begun to have a bird's-eye view of the universe from an unprecedented height. With the advent of remote sensing technology, people can more intuitively understand the changes of the earth every day. Among them, due to the rapid development of earth observation technology, it is possible to obtain remote sensing images of different time phases in the same area. The use of multi-temporal remot...

AI Technical Summary

Problems solved by technology

However, the disadvantage of this method is that the calculation of the vector difference value is directly performed on the multi-temporal remote sensing image, which is susceptible to interference from different noises in the image and different shooting angles of the remote sensing platform, resulting in a low detection rate.

However, the shortcomings of this method are: the method based on traditional supervised classification requires a large number of experts to spend time and effort to mark the ground features, which makes the method have certain limitations in promotion; at the same time, because the selection of the change threshold depends on manual selection or other clustering methods, making the development of supervised change detection methods limited

Images