This chapter describes six colorization methods that combine multiple images and their application to multispectral images that enhance night vision (NV). Based on multispectral image fusion and NV enhancement, the goal of colorization is to make the combined image resemble a daylight color picture for human decision making. Multispectral images enable robust NV object assessment in day/night conditions. Furthermore, colorized multispectral NV imagery can enhance human vision by improving observer object classification and reaction times, especially for low-light conditions. Qualitative (subjective metric) and quantitative (objective metric, such as the objective evaluation index) evaluations are applicable to the colorized images, which will be discussed in Chapters 9 and 10.

The six colorization methods include one color-fusion method and five color-mapping methods. Color fusion directly combines multispectral NV images into a color-version image by mixing pixel intensities at different color planes. Color mapping usually maps the color properties of a false-colored NV image (source) onto that of a true-color daylight picture (target). A color-mapping process typically consists of three steps: false coloring, color mapping, and contrast smoothing. Both pyramid- and wavelet-based multiscale fusion are utilized in the contrast smoothing (Chapter 7). The five color-mapping methods are segmentation-based mapping, statistical matching, histogram matching, joint histogram matching, and lookup table (LUT). Segmentation-based color mapping was introduced as a local coloring method in contrast with the global coloring method (i.e., statistical matching), whereas the joint-histogram matching (JHM) was a recently developed color-mapping method. The experimental NV imagery presented in this chapter includes visible (red/green/blue), image-intensified, NIR, and LWIR images.