Fusion of WorldView2 images using Contourlet, Curvelet and Ridgelet transforms for edge enhancement

Abstract

This article discusses the implementation of three transforms, namely Contourlet, Curvelet and Ridgelet, which are intended for image edge enhancement. These transforms were applied to fused Worldview 2 satellite images. The fusion was performed over the WorldView 2 satellite images applying various types of wavelet transforms, such as Daubechies, Bior, rbior, Coiflet and Symlet 5, with different levels of decomposition. The best results were obtained for the case of Symlet 5, level 5. Wavelet fused images and the generated images (using Contourlet, Curvelet and transformed Ridgelet) were evaluated and analyzed quantitatively. Quantitative methods in the present analysis include ERGAS, RASE, Universal Quality (Qu) and correlation coefficient (CC). Image merging and the implementation of transforms were performed with MatLab®, which supplies the following tools: Wavelet toolbox, Image processing toolbox, Contourlet toolbox, and Curvelet and Ridgelet source code. The results show that the Curvelet and Ridgelet transforms yield better results in terms of edge enhancement for both the merged image and the original image.