Natural-appearance Colorization and Enhancement for the Low-light-level Night Vision Imaging

The output video of the low-light-level solid-state imaging devices are always gray. For better low-light-level imaging applications, a natural-appearance colorization and enhancement method named Luminance Stretching Color Transfer (LSCT) for grayscale video images using color transfer is proposed. A two-channel natural-appearance color fusion method is refered to in the LSCT method. In order to achieve the natural-appearance colorization and enhancement, firstly, the pre-colorized image is obtained by combining the grayscale image with its negative image. Following this, an adaptive luminance stretching is performed and color of the reference image is transferred in the YUV color space. As compared with other methods based on color transfer, the LSCT method is less affected by the degree of similarity between the reference image and the original grayscale image. It means that relatively good results may be achieved for most scenes with an appropriate reference image. Thus, the LSCT method has better environmental adaptability. The comparisons reveal that the LSCT method is high efficient and its colorized results appear more natural in respect to human perception with better contrast and color harmony. Moreover, the LSCT method has been implemented in real time on hardware platforms. Therefore, it can effectively improve the effect of human observation to apply our method in the low-light-level imaging without increasing any hardware costs.