Compact Multispectral Camera Using RGB LED and Optimization

Multispectral camera gets the three-dimensional spatial information and spectral reflection of natural objects. It is useful to exploit the shape or color difference, even the material variation. This paper introduces a compact multispectral camera. It is composed by a gray camera and RGB LED. The multispectral images are recovered using quadratic optimization and known eigenvector of Mussel color chips. It is an active illumination method and the spectral response of LED and detector are calibrated to obtain the actual spectral reflection. We analyze the accuracy of recovered spectrum using simulation and build a prototype to validate it. The experiment shows that it can distinguish the close colors effectively. This multispectral camera can be an enhanced camera for intelligent detection.