Authors : Saad M. Darwish and Maha M. Ghoneim
Abstract: Collecting information from sensors with different physical characteristics increases the understanding of our surroundings and can provide 3D view. Moreover, 3D color images provide geoscientists, environment planners, mapping experts and military officers with easy-to-understand color images and useful height information which improves the interpretation of the environment of the areas of interest. This study proposes a novel method to optimize F-transform fused 3D images with feature matching technique based on Scale Invariant Feature Transform (SIFT) followed by Speed Up Robust Feature (SURF). Quantitative and visual results show that a more focused and cleared fused image is obtained after applying feature matching with SIFT followed by further refinement with SURF. The proposed method is robust and independent of scale, light intensity and orientation of camera. It shows that, F-transform is a promising 3-D multisensor image fusion algorithm that surpasses the previous approaches based on Hermite and wavelet transform due to its computational simplicity.
Saad M. Darwish and Maha M. Ghoneim, 2018. Optimization of 3D Fused Image Using Feature Matching Based on SIFT and SURF. Journal of Engineering and Applied Sciences, 13: 10133-10140.