Intelligent Images Analysis in GIS

The paper proposes an analysis of conventional pattern recognition methods and their use for images’ analysis in GIS applications. The methods like cluster analysis, neural networks, and immunocomputing are studied in detail. Classic methods of raster segmentation, including lines detection, levels’ difference, connecting of contours, threshold processing, and other are considered as applied to snapshots’ analysis. Special attention is paid to immunicomputing method implementation. The above method efficiency as well as its application’s domain is demonstrated with specific examples. Methods of isomorphism and metaclasses are developed to automatically recognize complex objects. The methods’ efficiency as well as advantages and disadvantages are manifested based on a real snapshot.