Infrared and Color Visible Images Fusion Based on Second Generation Curvelet Transform

Focusing on the fusion problem of the infrared image and the color visible image of the same scene, a novel image fusion algorithm based on the IHS transform and the curvelet transform is presented in this paper. Firstly, the color visible image is mapped from the RGB color space into the IHS color space and then the intensity component is attained. Secondly, the intensity component and the infrared image are merged based on the curvelet transform. Lastly, the fused color image is attained by applying the inverse IHS transform. When merging the intensity component and the infrared image, we discuss the fusion rules in detail. For the low frequency coefficients, we present a ´weighted averaging´ scheme based on the different physical features of the infrared image and the visible image; for the bandpass directional subband coefficients, according to the human visual characteristics, we present a selection principle based on the local area energy. Experimental results demonstrate the proposed algorithm not only can make the fused color image maintain the high resolution and the nature color of the visible image, but also can incorporate the detection and the unambiguous localization of the target, represented in the thermal image, as well.