Data Fusion Method for Multi-sources Remotely Sensed Imagery Based on IHS Transformation and Wavelet Analysis

Data fusion for multi-sources remotely sensed images is becoming a key-question in RS application. In this paper, two images differing in spatial resolution and spectral resolution are used in data fusion. The additive fusion technique based on IHS and wavelet transform is discussed, and an improved additive fusion technique is put forward. Firstly, the high frequency from high spatial resolution image after wavelet transform is added to the high frequency of the intensity from multi-spectral image after IHS transform and wavelet transform, which will get a integration of high frequency for a new intensity I´, while maintaining the low frequency of Intensity from the multi-spectral image as the low frequency for I´; secondly I´ is acquired by making the inverse wavelet transform; finally the fusion result is gained by making the inverse I´HS transform. The fusion test proves that the improved additive fusion technique is much more accurate in maintaining the characteristics of the multi-spectral image as well as improving the spatial resolution.