An effective color image fusion algorithm of visual and infrared night images

Abstract

An effective color image fusion algorithm based on NSCT(Nonsubsampled contourlet transform) and Lαβ color space is proposed for visual and infrared night image fusion, aiming at traditional fusion method suffering from no rich details and no clear target. The NSCT is applied in fusion processing because of its good properties which include multi-scale, time-frequency localization, shift invariance, and multiple directions discrimination, in order to obtaining exact position and clear image presentation of fused image. The Lαβ color space is applied to color transferring because of its no correlation. The oversaturated color of the fused image is overcome by adjusting the dynamic range of luminance component of Lαβ space. And the color of the fused image is more uniform, more natural. The experimental results show that the proposed method is effective.