Implementation of clustering methods for studying geodata arrays

Unprecedented big size and high dimension of existing sets of geodata make it difficult to reveal complex images. Clustering — one of the most important techniques for detection of geographical knowledge. However, existing methods for clustering have two serious lacks. First, valid methods for multiple parameter clustering have the limited capacity in recognition of spatial images. Secondly, available methods for clustering are not adapted absolutely for man-machine interaction. In work the approaches improving a clustering process for effective research of great and multivariate geographical data are analyzed. Special attention is given to the approach based on formation of hierarchical spatial clustering structures for detection of combined spatial images by using computing clustering methods.