QUALITY ANALYSIS OF 3D BUILDING MODELS FROM BASIC GEOSPATIAL INFORMATION DATA

• Main Author: Hidayah, Nurul
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/67609
• Institution: Institut Teknologi Bandung

3D city model is one of the recent alternatives that can be applied to help managing the urban areas in order to solve the complexity of urban area problems. Various applications can be done with 3D city models encouraging big cities in Indonesia to build their own 3D models. However, the development of 3D city models in Indonesia is often limited by relatively high production costs.One alternative that can be used to build a 3D city model is to use a large-scale Indonesian topographic map (RBI) that is already available for urban areas. RBI maps have the height attached to their data so they are referred as 2.5D data. The data can be used to build a 3D model by converting the building height into the building model height. This study will identify the problems that cause errors and quality tests from a 3D model that could be generated from an RBI map. The research process is carried out by selecting building feature on the RBI map. The building model height is obtained from the height of the building feature and the DTM height. This step was done by doing a quality test for the DTM, height calculation for 306.607 buildings in Bogor city, identifying the source of errors and quality test for the 3D model. The DTM quality test resulted in an RMSE value of 0.585 m and a LE90 of 0.966, which means it complies with the requirements needed for basic map class I. Errors that occur in the 3D modeling process consist of systematic, random, and blunders. Systematic errors are caused by DTM errors, random errors are caused by DTM data and points used for quality tests, and blunder errors are caused by the gaps in between DTM data and digitization errors. The quality test of the building model height produces a standard deviation of 0.451 m and an RMSE of 0.455 m. The geometric quality of the 3D building model from the height data and DTM data produces a standard deviation value of 1,022 m, RMSE of 1,040 m, and LE90 of 1,717.