Text
Building detection from coarse airbone laser scanner data
In developing countries, urban areas have high building densities with complex building pattern. Therefore, detecting and classifying building data can be challenging as it is difficults to separate between the building and non-building area. The buildings object from coarse ALS data which have noisy content and complexity of urban area were obtained by using information provided by topographic map. The FCN has been used to accomplish this research. The planar segementation has been conducted to detec the roof of the buildings object. The spatial information from 2D map was used to observe the location and the shape of the building. The Spatial information from 2D map help to easily recognize the variation shape of the building. The points clipped using polygon from 2D map data to assign points that belong to building objects.rnThe trained network was tested into different region in Indonesia. The accuracy assessment calculated based on the F-Score. FRom RMSD in LOmbok region, the cleaning data perform better than the noisy data with RMSD equel 1.03%. The number of sites enchance the performance of the classification. The RMSD in another region that in Tanjung Lesung 20.33%, Tanggamus 30.5% and MAkasar 19.70%.
B20190424293 | 621.3678 INT b | Perpustakaan BIG (600) | Tersedia |
Tidak tersedia versi lain