Parallel Algorithm of Area Object Boundaries Identification

P.A. Kim (Russia)

Keywords

Modelling, multiprocessing, image processing, geoinformation system.

Abstract

The task of choice and allocation of the area region's objects on the image, is well-known and in detail investigated, so on the basis of the developed theoretical recommendations there are also many programming decisions usually used in various spheres of a national economy: geology, a forestry, ecological monitoring etc. The transfering of existed and known algorithms and programs at the consecutive computers into the parallel computing architecture and the distributed computational structures for data processing becames the very actual problem. Geoinformation systems deal with allocation of the area objects at the aero-space images. The parallel analysis of area regions appears essentially more difficult, and results in necessity of identification of the objects. The mathematical aspects of the identification problem may be expressed in theorems of existence and uniqueness of identifying attributes for the given type of Decart's metrics. The work was carried out partly with financial sponsoring of projects RFBR (№ 05-07-90057- в, № 05-05-98006-р_обь_а, № 05-07-98011-р_обь_в.).

Important Links:



Go Back