Information losses minimization in the procedure of images multispectral processing using acousto-optic diffraction elements

Multispectral processing is strong and useful method which produces information regarding spatial and spectral composition of images to be studied. However, as any other physical operation, multispectral processing causes information losses which can be distributed by certain way between spectral, spatial, and gradation kinds of information. The losses of this information are caused by the noise of fluctuation nature. Hence, in order to minimize the information losses, it is necessary to find the source of noise which exercises the decisive influence on the signal-to-noise ratio. Also the information structure of the device providing the images multispectral processing has been considered. Special attention is paid to the features of diffraction elements – acousto-optic tunable filters – which have very specific information structure and which improvement can provide information losses minimization in the system of multispectral images processing.