Image fusion using D-S evidence theory and ANOVA method

Abstract

Image fusion is an important embranchment in information fusion. In image processing, information fusion techniques are able to significantly reduce uncertainty and inaccuracy in the information obtained from any single source alone. In this paper, a new method based on D-S evidence theory and ANOVA (Analysis Of Variance) method is proposed for image fusion. ANOVA method is employed for detecting potential edges in image. To deal with the uncertain weak edge and the difficulty in threshold selection, D-S evidence theory is then applied. Since D-S evidence theory has the advantage of conveniently representing uncertain information and dealing with information from multi-images, edges to be fused could be preserved as much as possible, which may be useful for further processing, such as image analysis and image understanding. The proposed image fusion method can be applied to fusing different types of images, such as visible and infrared images, remote sensing images and so on. Experiment results on the fusion of different types of images have demonstrated the robustness and efficiency of the proposed method. It has been shown that the fused edge image has more complete and reliable edge information than that from any of the original image.