Target detection threshold in noisy color imagery

Current target acquisition models are for monochrome imagery systems (single detector). The increasing interest in multi-spectral infrared systems and color daylight imagers highlights the need for models that describe the target acquisition process for color systems (2 or more detectors). This study investigates the detection of simple color targets in a noisy color background. Color targets are varied separately either in hue or saturation. Noise is created with a mixture of random hue and saturation combinations. Our preliminary result showed a simple two-color (yellow-blue) representation did not improve the standard black-and-white Minimum Resolvable Temperature Difference sensitivity. Subsequent psychophysical experiments reveal that human hue and saturation discriminations interact (the Abney effect) and need to be separated in color target detection modeling. Research is continuing to better define the mathematical relationship between the target acquisition parameters (e.g., temperature difference or intensity contrast) and the color space of hue and saturation.