Using Machine Learning to Extract Building Inventory Information Based on LiDAR Data

The extraction of building inventory information is vital for damage assessment and planning modelling studies. In the last few years, conventional data was overcome using various remote sensing techniques. main objectives this study were to supply necessary structural engineers calculate seismic performance existing structures. Thus, we investigated light detection ranging (LiDAR) derivatives classify buildings extract information, such as different heights footprint area. most important achieve also classified machine learning methods, Random Forest, Tree, Optimized over object-based segmentation results. All methods successfully with high accuracy, whereas other outperformed RT. height area results show that archived sensitivity sufficient be further used in applications, detailed health monitoring. Overall, presents a methodology can accurately information. results, future studies directed investigations on determining construction year data, multi-temporal satellite imagery.