UDC 004.021
MODERNIZED ALGORITHM OF CLUSTERING WITH KOHONEN NEURAL MAP IN THE PROBLEM OF PROCESSING HYPERSPECTRAL IMAGES
M. N. Guryeva, Senior Engineer, Sberbank-Technologies, Ryazan; This email address is being protected from spambots. You need JavaScript enabled to view it.
The problem of improving the results of clustering a hyperspectral image for the possibility of subsequent identification of objects is considered. As a method of optimal clustering, a neural map of Kohonen with an automatic determination of object class number is presented. The results of the influence of several types of filters on the quality of clustering are presented. The aim of the work is to find the optimal method for pre-filtering channel plane section of a threedimensional array of data. Quality indicators include the speed of presetting and building a map, the number of automatically detected groups of objects.
Key words: clustering, hyperspectral images, neural map, filtration, remote Earth sensing.