V.V. Shapovalov, B.S. Gurevich, S.V. Andreyev, A.V. Belyaev, V.N. Chelak

Spectrophotometers intended for the biological objects study, include the important unit which is responsible for separation of the spectral lines at which the absorption or reflection from the objects are measured. As for the imaging spectrophotome-ters, such units perform the operation of the monochrome subimages selection from the total polychromic image of the ob-ject. In the devices which are the subject of our consideration, the device which is responsible for spectral selection, is acousto-optic tunable filter (AOTF). Its advances are caused by the electronic (including programmable) way of the selected spectral line control. It allows to avoid the mechanically moving parts, which provides high operation rate and reliability. Due to this, the study of such spectrophotometers spectral selectivity means the study of AOTF spectral selectivity. Pro-ceeding from the fact that the spectral resolving power criterion which is the subject of further consideration, is based on the information principle, it is necessary, before this criterion formulation, to consider the basic information characteristics of AOTF and to establich their connection with the parameters of acousto-optic interaction. Such information characteristics as information capacity and information transmission capability are considered. The nature of AOTF spectral selectivity is studied using the statements of the Kogelnik-s coupled waves theory. The information criterion of spectral resolving power is proposed, according to which the two spectral lines are resolved if in each of them it is possible to recognize the given amount of gray scale levels with the probability njt less than the given value. According to the proposed criterion, the method of AOTF spectral resolving power measurements has been elaborated. The commonly accepted Raleigh criterion is often inapplicable for estimation of both spatial and spectral resolution of opt-ical devices, including spectral devices. The proposed information criterion allows to avoid some deficiencies of Raleigh criterion which is connected with the fact that the proposed criterion takes into consideration both gray scale levels which must be recognized in each resolved elements (spectral or spatial) and the transmitted signal noisiness.