Infrared and Color Visible Image Sequence Fusion Based on Statistical Model and Image Enhancement

Abstract

A novel fusion method is proposed for image sequence which based on the non-Gaussian statistical modeling of wavelet coefficients and image enhancement in IHS color space of color visible image. Firstly, the original color visible image is transformed into a perceptually decorrelated color space in order to treat the achromatic and chromatic components separately. Then, the achromatic component and infrared image are combined by a statistical fusion method based on dual tree complex wavelet transform (DT-CWT) using the generalized Gaussian distribution (GGD). The means and variances between the fused component and the original achromatic component are matched by a linear remapping for image enhancement. At last, the color space is transformed back into the RGB color space. This method not only have a superior fusion performance but also can effectively produce a high-contrast color fused image with the similar natural characteristics as the original color visible image.