No abstract

No abstract

No abstract

Results of computational experiments aimed at studying the accuracy characteristics of interchannel gradient reconstruction of signals of color digital images distorted by impulse noise with different brightness distribution laws, which are not correlated between the channels, are generalized. Examples of interchannel reconstruction of color digital images on the basis of different sets of gradient estimates are presented.

An approach to the construction of a hyperspectral system is proposed and justified, providing control of spectral, spatial, and radiometric (brightness) resolution, which opens up the possibility of hyperspectral monitoring of objects with limited computing resources and bandwidth of the video data transmission channel. Spectral resolution control is supposed to be carried out on the basis of tuneable, software-controlled acousto-optic filters, spatial resolution control – based on modern CMOS matrix technologies with the “active pixel” function, digitalization control of the images obtained – based on the developed analogy-to-digital converter with tuneable quantization level. Modelling using experimental data has shown the possibility of implementing a sufficiently reproducible classification of hyperspectral images with a decrease in their spectral, spatial, and radiometric resolutions.