摘要
多源遥感影像匹配后质量变低的问题一直困扰着遥感信息提取及应用,探索一种基于自适应遗传算法的影像信息分层匹配,以归一化积来度量两幅遥感影像的相似性,并用该方法求解最高适应度,实现上层初匹配,然后使用仿射变换模型变换初匹配后的多源遥感影像,通过优化粒子群算法优化变换后多源遥感影像中的区域互信息,最终实现多源遥感影像信息下层精细匹配。以SAR数据与光谱影像的实验证明,该方法能显著提升匹配后的遥感影像,并能实现多源遥感影像的分层精准匹配。
The problem of low quality of multi-source remote sensing images after matching has always plagued the extraction and application of remote sensing information.To explore a layered matching of image information based on adaptive genetic algorithm,the normalized product is used to measure the similarity of two remote sensing images,and this method is used to solve the highest fitness to realize the initial matching of the upper layer.Then the affine transformation model is applied to transform the multi-source remote sensing image after initial matching.The regional mutual information in the transformed multi-source remote sensing image is optimized by the particle swarm algorithm,and finally it is realizedfor the fine matching of lower layer of multi-source remote sensing image information.The experiment on SAR data and spectral images proves that this method can significantly improve the matched remote sensing images,and can achieve hierarchical and accurate matching of multi-source remote sensing images.
作者
方攀
贾睿洁
高娟
FANG Pan;JIA Ruijie;GAO Juan(The Third Geological and Mineral Survey Institute,Henan Geological and Mineral Exploration and Development Bureau,Xinyang 464000,Henan,China;Henan Science and Technology Innovation Center of Natural Resources(Application Research of Information Perception Technology),Xinyang 464000,Henan,China)
出处
《矿产与地质》
2022年第6期1287-1292,共6页
Mineral Resources and Geology
基金
河南省地质矿产勘查开发局地质科研项目"遥感与无人机技术在矿山环境影响调查中的应用研究"(豫地矿文〔2021〕7号)资助。
关键词
自适应遗传
多源遥感
影像信息
分层匹配
归一化积
优化粒子群
adaptive genetics
multi-source remote sensing
image information
hierarchical matching
normalized product
optimized particle swarm
