Video Fusion Using Pixel Averaging, Principal Component Analysis and Laplacian Pyramid ? A Comparative Study

In this paper, performances of various video fusion algorithms are compared by applying them to a set of infrared (IR) and visible band videos. The application of interest is area surveillance and the fusion process aims at integration of complementary information from multi-sensor inputs for enhancing the human perception of the monitored scene and to make the result suitable for further processing. The performance of algorithms viz. pixel averaging, principal component analysis and Laplacian pyramid are compared. A set of measures of effectiveness for comparative performance analysis like Fusion Factor and Fusion Symmetry are defined and applied on the output of the above fusion algorithms.