NAUFAL SULISTYANTO, 22314072 (2026) EVALUASI PEMANFAATAN GEOAI UNTUK PEMODELAN TIGA DIMENSI BANGUNAN PADA WILAYAH PERKOTAAN DAN PERDESAAN. Diploma thesis, Politeknik Agraria STPN.
![[img]](http://repository.stpn.ac.id/style/images/fileicons/text.png) |
Text
Naufal Sulistyanto_22314072.pdf Download (506kB) |
Abstract
The utilization of GeoAI for three-dimensional building modeling represents a technological breakthrough with significant potential to accelerate the provision of land geospatial data. However, in practice, there is a gap between the application of GeoAI technology and the accuracy achieved in 3D building modeling, influenced by differences in spatial characteristics between urban and rural areas, as mandated in the Terms of Reference for Aerial Photography and LiDAR Mapping to enhance 3D baseline data in 2024. This study aims to examine the process of 3D building modeling using GeoAI in both urban and rural areas, and to evaluate the dimensional accuracy of the resulting 3D models against the land base map. This research employs a mixed-methods approach with a sequential exploratory design. The qualitative phase examines the 3D building modeling process utilizing GeoAI, analyzed using the Miles and Huberman model. The quantitative phase involves dimensional accuracy testing of the 3D models, with data analysis techniques including completeness testing, geometric accuracy testing, and validity testing on a sample of building distributions across two areas: the urban area of Gondang Village, Gondang Sub-district, and the rural area of Toyogo Village, Sambungmacan Sub-district, in Sragen Regency. The findings reveal that the 3D building modeling process using GeoAI consists of building footprint extraction (LoD 0), building height identification, and extrusion of the building footprint against the height data. The quantitative results, validated against the 2024 Terms of Reference for Aerial Photography and LiDAR Mapping, show the following: the completeness test for the 3D model in the urban area was 94.62%, while in the rural area it was 89.33%; the geometric accuracy test (RMSE) was 0.116 meters for the urban area and 6.549 meters for the rural area; and the validity test of dimensions, for all sampled building points in both areas, showed p-values above the 0.05 significance level. In conclusion, the utilization of GeoAI is effective and efficient for large-area 3D building modeling in both regions, with no statistically significant difference between the dimensions measured in the field and those represented by the 3D models produced using GeoAI technology. Recommendations include refining and repositioning building footprints to match actual ground positions, using supplementary LiDAR data in rural areas, conducting direct field measurements on buildings obscured by vegetation, and comparing SAM performance with other algorithms such as Mask R-CNN or YOLO. Keywords: GeoAI, 3D Building Modeling, Urban Areas, Rural Areas
| Item Type: | Thesis (Diploma) |
|---|---|
| Subjects: | H Social Sciences > HD Industries. Land use. Labor > HD29 Pertanahan |
| Divisions: | Prodi Diploma IV Pertanahan |
| Depositing User: | yosep ka perpus |
| Date Deposited: | 10 Sep 2026 09:25 |
| Last Modified: | 10 Sep 2026 09:25 |
| URI: | http://repository.stpn.ac.id/id/eprint/4960 |
Actions (login required)
 |
View Item |
