A new image fusion method based on multiscale analysis and color model

Contourlet, one of multiscale analysis tool with great property of multi-directionality, locality and sparse approximation, is an excellent 2-dimentional image description approach. Contourlet can be employed as a remote sensing image fusion approach to improve spatial resolution and application through precise capture of edge information from panchromatic image and injection of that into multispectral image. Hence, image fusion using contourlet has become the research highlight and fronts in the field of remote sensing image fusion. However, the unwanted spectral distortion appears after the contourlet method is applied to remotely sensed image. Lαβ color model is composed of 3 orthogonal color channels with least correlation, which are achromatic, yellow-blue and red-green ones. Spectral distortion can be remarkably reduced when the fusion is operated in only achromatic channel due to its non-colorful characteristic. Hence, lαβ-Contourlet fusion algorithm is proposed. Panchromatic image and multispectral images from BJ-1 are then used to test the united algorithm. Assessment indicates that lαβ-Contourlet fusion algorithm can improve spectral performance of fusion result simply from contourlet algorithm.