CBERS-02B星HR高分辨率遥感数据融合研究分析

在阐述遥感影像融合原理的基础上,以贵州省贵阳市部分区域为研究区,应用PCA,Brovey,SVR,IHS共4种融合方法对CBERS-02B星HR数据与CCD数据进行融合实验。从光谱信息和空间信息两个方面对融合效果进行分析,对比4种融合方法对CBERS-02B星影像数据的适用程度,同时与SPOT-5融合数据进行对比分析。研究表明:(1)4种融合方法中,SVR变换法最适合CBERS-02B星遥感数据的融合,融合数据的各评价指标均高于其他融合方法,空间信息丰富,图像细节效果良好;(2)CBERS-02B星同源遥感数据融合影像的质量低于SPOT-5影像融合数据;(3)CBERS-02B星CCD数据的质量低是导致融合影像质量不高的主要原因,所以在CBERS-02B星HR数据的使用过程中应尽量寻求其他来源的多光谱数据进行融合。

Study on Image Fusion for CBERS-02B HR Image

Based on principles of the RS image fusion,the HR data and CCD data of CBERS-02B were fused with the four image fusion methods: PCA,Brovey,SVR and IHS in the study area of part region of Guiyang in Guizhou Province.Based on the spectral information and spatial resolution perspective,the fusion results were analyzed,the adaptability of all the fusion methods for the CBERS-02B image fusion were studied and were compared and analyzed with the fused SPOT-5 data.The study results showed that:(1) among the four fusion methods,SVR was the most suitable for the CBERS-02B image fusion,the evaluation indices for fused data were higher than those of other fusion methods;it contained much more information and the detailed information showed better;(2) the image fusion quality of the congenetic CBERS-02B RS data was lower than that of the fused SPOT-5 data;(3) the low quality of CBERS-02B CCD image was the main cause of the poor fusion image,therefore,when the CBERS-02B HR data was used,the spectral data of other satellites should be sought to fuse.