A New Pixel Level Image Fusion Method based on Genetic Algorithm

Background/Objectives: To propose a new fusion technique for combining optical and IR images and validate the proposed technique with the existing techniques using entropy as an evaluating measure. Methods/Statistical Analysis: In this paper we propose a new pixel level fusion method using Continuous Genetic Algorithm (CGA) using Heuristiccrossover for reproduction. Findings: Pixel level Fusion methods are computationally less complex and converge quickly. The proposed approach is applied on multispectral images which are used in applications like multispectral face recognition, Medical imaging, Remote Sensing etc. The proposed algorithm requires less memory space and has less computational complexity. Conclusion/Improvements: An increase in the entropy of the fused image indicates that there is an increase in the overall information content. The proposed technique is implemented on a set of visual and thermal images and an increase in the entropy value of the fused image is observed.