Spectral pattern recognition using multi-band acousto-optic filters

The problem of detection of objects on their spectral patterns by hyperspectral mapping is considered. The main issues hindering an effective and fast solution of the problem are described. The method of on-line object–background discrimination is validated, which utilizes multi-band tunable spectral filters. The principles of operation and technical implementation of these devices are formulated. The assumptions forming the basis of this method and its features are discussed. An original method based on a modulated acousto-optic filter is proposed. We have made conclusions on a conceptual possibility of the considered methods based on a novel technique of multi-band acousto-optic filtering, on its relevance, a short-time response, operational efficiency, and adjustability.