S-Space College of Social Sciences (사회과학대학) Institute for Korean Regional Studies (국토문제연구소) 지리학논총 (Journal of Geography) 지리학논총 Volume 21/22 (1993)
다축척 수치 지도 구축을 위한 선형 사상의 일반화에 관한 연구
- Issue Date
- 서울대학교 국토문제연구소
- 지리학논총, Vol.21, pp. 17-33
- 다축척 수치 지도; 일반화; 단순화 지료; 계층적 자료구조; 형태적 대표점; multiscale digital map; line simplification; hierachical data structure; critical points; spatial accuracy
- Data incorporated into GIS are derived from many different sources, both analogue and digital. The primary source is still undoubtedly the hardcopy map. In developing geographical databases from maps, however, some problems are encountered. First, it is difficult to obtain the map data at appropriate scale to the analysis purpose. Second, in cast that the data of many details level are needed according to the purposes, spatial data for the same area have to be stored at a number of distinct scales. It generates the redundancy of spatial data. In this context, this study investigated the development of 'Multiscale Digital Map Routine (MDMR)' which extracted from relatively large scale data to smaller ones the users wanted to get. In order to simplify cartographic line features in accordance with scale reduction, MDMR used 'critical points search routine' created here, which idicated the local properties of line features, and Douglas-Peucker's algorithm. The major findings were summed up into the followings. First, it was found out that the hierachy of simplification index, which was the core in Cromley's method, was not always guranteed. Douglas-Peucker's algorithm. Cromley used as a simplification prodedure, simplified line features with global point of view. It was validated that the local property of line features in such global approach was apt to be neglected and the construction of hierachy with no thought of locality might entangle the hierachy. Second, MDrv1R designed new method, namely . critical points search routine; and then constructed the hierachical data structure based on these critical points. In consequence, MDMR corrected most entanglements, particularly entanglements of higher rank in hierachical structure.