The article analyzes the methods of data fusion partial spectral channels multispectral system of monitoring objects and scenes, working in various parts of the electromagnetic spectrum whose essence is to implement process optimal joint processing of reflected signals and self-radiation of the observed objects. Shown basic requirements are put before the task of interconnecting partial spectral channels.
It is proved that for maximum information content data obtained by remote monitoring is necessary to use sensors of various physical nature, which operate in different parts of the electromagnetic spectrum, namely microwave (RF), infrared and optical diapasons.
The analysis of the two methods of images data fusion, the first of which is based on increasing the differences between the images obtained from various partial channels multispectral monitoring system, and the second on the separate analysis of low-frequency and high-frequency spectral components of the image. Moreover, both of these algorithms complement each other and can be used simultaneously.
It was found that significant difficulties in the images data fusion discussed methods associated with the exact matching partial channels with spatial coordinates; search for best practices to eliminate the inverse display of individual objects in different zones of the spectral and radiometric distortion (noise); selecting the best algorithm for determining the weighting coefficients of the partial channels, and the like.
Described the idea of a possible implementation of the system of monitoring objects and scenes with maximum frequency diversity partial spectral channels and propose options for the overall structure of a multispectral system.