A Physics-based Fusion Approach for Color Night Vision

Common night vision systems use single-band sensors that are sensitive to either emitted or reflected infrared (IR) radiation, and thus provide a monochromatic image of the inspected scene. Adding color can increase depth, contrast and perception, improving soldiers' reaction time and reducing fatigue. However, current color night vision systems are quite limited in performance and range of application. In Phase I, Spectral Sciences, Inc. (SSI) demonstrated the feasibility of a novel physics-based approach involving fusion of multispectral thermal IR and near-IR imagery to provide natural-looking color rendering with little or no prior knowledge of the scene required. SSI proposes further development, testing, and validation of the approach in Phase II. Thermal imagery from both MWIR and LWIR spectral regions will be investigated. To assist algorithm development, testing and validation and build a color knowledge base, SSI will team with Physical Sciences, Inc. in a field measurement program to collect simultaneous hyperspectral LWIR, MWIR and broadband near-IR imagery of suburban and rural scenes over a wide range of environmental conditions. Results of Phase II will include the color rendering algorithm, measurement database, and a conceptual design for a multispectral IR night vision system.