(A) study on a micorcomputer based image processing system for remote sensing

Abstract

A microcomputer based image processing system, named KAIS-MIPS, is successfully developed for the purpose of the processing and analysis of remotely sensed Landsat data. This system is found to be very advantageous in cost, flexibility, and accessibility in comparison with large computer based image processing system. In this thesis, first, various algorithms for Landsat data processing implemented in KAIS-MIPS are presented. Then, using KAIS-MIPS, two kinds of land use classification of Inchun and vincinity in Korea are completed by unsupervised and supervised analysis, and the resultant maps are produced. Finally, the classification performances in both cases are evaluated by 55 test sample areas obtained from 1:50000 map.