INFRARED AND VISIBLE IMAGE FUSION USING HYBRID LWT AND PCA METHOD

Abstract

Image fusion is a method in which all the relevant information is collected from the input source images and included in few/single output image. Image fusion techniques are divided into two broad categories: spatial domain and transform domain. Principal component analysis (PCA) is a spatial domain technique which is computationally simpler and reduces redundant information but has the demerit of spectral degradation. Lifting wavelet transform (LWT) is a transform domain technique which has an adaptive design and demands less memory. In this project, a novel hybrid fusion algorithm has been introduced which combines the LWT and PCA in a parallel manner. These two fusion methods are applied on Infrared and Visible image data set. Infrared and visible images contain complementary information and their fusion gives us an output image which is more informative than the individual source images. The hybrid method is also compared with conventional fusion techniques like PCA, LWT and DWT. It has been shown that the proposed method outperforms the conventional methods. The results are analyzed using performance

parameters standard deviation, average value, the average difference, and normalized cross- correlation.