Multiresolution region-based image fusion using the Contourlet Transform

Different sensors provide a variety of images with different specifications (spectral, spatial and radiometric resolution, etc.). Image fusion techniques have been utilized to benefit the best features of all input images and to provide better application-wise output images. In this paper, a new region-based image fusion technique using the Contourlet Transform (CT) is proposed to produce a fused image better for human and machine interpretation and to reduce the computational effort of the traditional techniques. Due to the high directionality and anisotropy of the CT, the proposed technique is mainly developed to solve the problem of capturing the fine lines and contours of the input images. In this technique, the input images are segmented into small regions more suitable for the proposed fusion approach, where the segmentation process is performed in the frequency domain for better results. The fusion decision is made based on a new quality assessment scheme for each segmented region. Also, the presented region-based fusion approach is more robust than the traditional pixel-based techniques, where it reduces: the blurring effects, sensitivity to the misregistration, and noise effect in remote sensing images.