APPLICATION OF EXPERT SYSTEM AND IMAGE PROCESSING TECHNIQUES FOR DETECTING BUILDING 
CHANGES IN GEO-SPATIAL DATABASES 
Prof. Bock-Mo Yeu 
Dept. of Civil Engineering, Yonsei University 
Prof. Duk-Jae Sohn 
Dept. of Civil Engineering, Daejin University 
Prof. Hwan-Hee Yoo, 
Dept. of Urban Engineering, Gyeongsang National University 
Senior Researcher Jae-Hong Yom 
Geographic Information Research Institute, Hanjin GIS Co. 
Commission IV, Working Group 3 
KEY WORDS: Urban, Change_Detection, Expert_System, Building_Changes, Geo_Spatial Databases 
ABSTRACT 
Researches are active in the field of photogrammetry to automate interpretation of cartographic objects. An attempt is 
made here to design an automated system which could be applied to practical municipality administration of handling 
geo-spatial data. Existence of past vector maps in digital format is assumed which is used as a priori knowledge as 
well as to confirm visually the result of the automated detection. 
1. INTRODUCTION 
Detecting new buildings and changes to existing 
buildings in urban areas to check for compliance with 
regulations and permits, is one of the responsibilities of 
all urban municipalities in Korea. Detecting these 
changes is being carried out manually by interpretation 
of aerial photos and comparing to past maps of the same 
area. Efforts have been made to automate this process by 
applying image processing techniques used in remote 
sensing such as producing image ratio maps and 
difference maps of the same area acquired at different 
points of time. But detecting annual building changes 
involve large scale aerial photos and involves 
interpretation of details which are not available in the 
smaller scale satellite imageries. 
Manual photo interpretation is carried in local 
municipalities to identify new buildings and changes in 
building structure by means of aerial photograph. 
Identified changes are then marked onto a paper map 
988 
and to a paper form. The form sheet is then used as a 
reference material to check for illegal buildings and take 
necessary measures. The changes are classified into 
general area and development restriction area. The types 
of changes for general area is divided into new 
construction, extension of building, remodeling of 
building. The types of changes for development 
restriction area is divided into 7 different types, new 
construction, extension of building, remodeling of 
building, landuse change, cutting woodland and others. 
The main problem of the present method of handling 
photo interpretation is that it is time consuming for 
timely update of the building records. Also lack of 
manpower for photo interpretation poses a problem as 
the changes in cities get faster and larger in terms of 
spatial extent.(Yeu, 1994) 
Image segmentation can be applied to large scale 
scanned aerial photos and the extracted features are then 
labeled and values for various descriptors are computed 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B4. Vienna 1996