Qualitative and quantitative comparisons of multispectral night vision colorization techniques

Abstract

Abstract. Multispectral images enable robust night vision (NV) object assessment over day-night conditions. Furthermore, colorized multispectral NV images can enhance human vision by improving observer object classification and reaction times, especially for low light conditions. NV colorization techniques can produce the colorized images that closely resemble natural scenes. Qualitative (subjective) and quantitative (objective) comparisons of NV colorization techniques proposed in the past decade are made and two categories of coloring methods, color fusion and color mapping, are discussed and compared. Color fusion directly combines multispectral NV images into a color-version image by mixing pixel intensities at different color planes, of which a channel-based color fusion method is reviewed. Color-mapping usually maps the color properties of a false-colored NV image (source) onto that of a truecolor daylight target picture (reference). Four coloring-mapping methods-statistical matching, histogram matching, joint histogram matching, and look-up table (LUT)-are presented and compared, including a new color-mapping method called joint-histogram matching (JHM). The experimental NV imagery includes visible (Red-Green-Blue), image-intensified, near infrared, and long-wave infrared images. The qualitative evaluations are conducted by visual inspections of the colorized images, whereas the quantitative evaluations are achieved by a newly proposed metric, objective evaluation index. From the experimental results according to both qualitative and quantitative evaluations, the following conclusions can be drawn: the segmentation-based colorization method produces very impressive and realistic colors but requires intense computations; color fusion and LUT-based methods run very fast but with less realistic results; the statistic-matching method always provides acceptable results; histogram matching and joint-histogram matching can generate impressive and vivid colors when the color distributions between source and target are similar; and the statistic-matching then joint-histogram matching (SM-JHM) method is a reliable and efficient method recommended from both qualitative and quantitative evaluations.

Contextual Tracking Approaches in Information Fusion

“…Context awareness including: environments, sensors, and targets, improved multi-source robustness [98] and clutter suppression [99]. Currently, context-enhancement is being mapped to common qualitative and quantitative results for user assessment (DFIG Level 5) and refinement of context moving and stationary target data (DFIG Level 1) [100]. Next we present an overview of the categories of contextual analysis for target tracking based on the literature review.…”

Improved Color Mapping Methods for Multiband Nighttime Image Fusion

“…The second full-reference metric is the color natural metric (CNM: [78,81,82]). The CNM measures the similarity of the color distributions of a color fused image and a daylight reference color image in Lab color space using Ma's [83] gray relational coefficients.…”

“…The third full-reference metric is the objective evaluation index (OEI: [81,82]). The OEI measures the degree of correspondence between a color fused image and a daylight reference color image by effectively integrating four established image quality metrics in CIELAB L*a*b* color space: phase congruency (representing local image structure; [84]), gradient magnitude (measuring local image contrast or sharpness), image contrast (ICM), and color naturalness (CNM).…”

Information fusion reliability analysis for component survivability

“…Examples of information fusion include tracking accuracy [37,38], tracking filter credibility [39], and object detection credibility [40,41] which are important for information quality and quality of service metrics [42].…”