Colorization of infrared images based on DWT fusion and color transfer

The conventional means to colorize an infrared image are pseudo-color coding methods, which build the mapping relations between grey-level and color model artificially. These methods have not taken the natural color of the infrared image into account. As the development of the method of color transfer between images in recent years, the natural colors of the gray-scaled image can be obtained to a considerable degree by using color transfer technique. But this method is mostly applicable to the gray-scaled image with high-contrast and good textural features. As the infrared images have the characteristic of low-contrast and bad textural features, they cannot adopt this method directly. This paper takes the advantage of the complementation information of the infrared image and low-level light image, first fuses the two images by using discrete wavelet transform (DWT) fusion to get an image with high-contrast and good textural features, and then colorizes the image with the color transfer method. Experimental results show that the image after color transfer resembles the natural image, reflects the details of the both images, and outperforms the conventional pseudo-color coding method greatly.