An Improved SVM+GA Relevance Feedback Model in the Remote Sensing Image Change Information Retrieval

With the rapid development of satellite remote sensing technology, the volume of image datasets in many application areas is growing exponentially and the demand for Land-Cover and Land-Use change remote sensing data is growing rapidly. The full complexity of the scenes and the interactive retrieval for the change information has become increasingly difficult. To address these challenges, this study proposed an improved SVM+GA relevance feedback model in the remote sensing image change information retrieval. The proposed approach can take consideration of avoiding local maxima in the SVM kernel parameters optimization and the subset feature selection simultaneously by combining GA. The remote sensing image change information retrieval experimental results compared with the remote sensing image change information retrieval model showed that the proposed relevance feedback approach can achieve a promising result.