Open AccessСпектральное распознавание объектов с помощью многооконных акустооптических фильтров
Author(s) -
В.Э. Пожар,
Д.Ю. Великовский
Publication year - 2020
Publication title -
žurnal tehničeskoj fiziki
Language(s) - English
Resource type - Journals
eISSN - 1726-748X
pISSN - 0044-4642
DOI - 10.21883/os.2020.07.49578.107-20
Subject(s) - computer science , filter (signal processing) , hyperspectral imaging , basis (linear algebra) , object (grammar) , optical filter , filtration (mathematics) , electronic engineering , computer vision , artificial intelligence , optics , engineering , mathematics , physics , geometry , statistics
The problem of object detection using their spectral characteristics is considered. The basic problems, preventing to prompt and effective solution, are described. A method for real-time objects discrimination from background in hyperspectral monitoring is presented, which uses multi-band tunable spectral filters. Operation principles and technical implementation of these devices are specified. The novel implementation basis is described using a modulated acousto-optical filter. The assumptions, which support the method, are discussed. It is concluded that the described methods based on multi-band acousto-optical filtration technique is realizable and promising, and can became an effective and flexible tool for express-analysis.