Study on fusion methods of ASAR and ETM+ data and information extraction

Abstract

Information extracted from remote sensing data plays an important role in the environment, hydrology and geology study. Optic remote sensing image has plenty of spectrum information and microwave can reflect land surface texture and penetrate ground to some extent. Fusion of microwave and optic remote sensing image will take advantage of mutual complementary information, and extract subsurface information more available. A comprehensive fusion approach between different remote sensing data was proposed, and the ASAR and ETM+ data were chosen as data source. Firstly, panchromatic and multi-spectral images of ETM+ were fused with principal component analysis (PCA) method. Spectral information and spatial detail information of the merged image has been enhanced compared to the original images. Secondly, ASAR and merged ETM+ data were fused using three methods, including multiplicative, Gram-Schmidt and discrete wavelet transformation (DWT). DWT fusion was the primary research content. The image quality after fusion was evaluated by means of visual effects, entropy, average gradient, correlation coefficient and standard deviation. The results show that the image fused with DWT has the highest accuracy, in which more surface and subsurface information can be expressed better. This research will build a foundation for making full use of ASAR and ETM+ data.