Hryvachevskyi A. Improving the informativeness of multispectral monitoring systems by image fusion of the visible and infrared ranges.

In the dissertation, an actual scientific and technical task was solved for the development of effective methods for increasing efficiency, namely, raising the information content and reducing the amount of redundant information in multispectral optical-electronic systems for monitoring objects and scenes that incorporating imaging sensors of the visible and infrared ranges of electromagnetic waves. The multi-criteria metric for evaluating the informativeness of images of unknown plot content is improved. The developed method and algorithm for multimodal integration of the visible and infrared ranges images based on discrete wavelet transform, in which, unlike the known methods, the formation of low-frequency wavelet coefficients of the resulting wavelet spectrum of a complex image is based on the regression analysis of the corresponding parts of the incoming wavelet spectra, informativeness of the fused image is improved by about 10%, as well as reduced the amount of data without useful information on the average by 26.72%, compared with the amount of data received on partial channels of the visible and infrared ranges. An adaptive method and algorithm for detecting moving objects to destabilizing factors during video monitoring are developed. Based on the developed technologies, it was possible to increase the probability of moving objects detecting, recognizing and identifying with a multispectral monitoring system, which in turn made it possible to increase the effective observation distance up to 11%.