A 3D Shape Descriptor Based on Contour Clusters for Damaged Roof Detection Using Airborne LiDAR Point Clouds

Abstract

Abstract:The rapid and accurate assessment of building damage states using only post-event remote sensing data is critical when performing loss estimation in earthquake emergency response. Damaged roof detection is one of the most efficient methods of assessing building damage. In particular, airborne LiDAR is often used to detect roofs damaged by earthquakes, especially for certain damage types, due to its ability to rapidly acquire accurate 3D information on individual roofs. Earthquake-induced roof damages are categorized into surface damages and structural damages based on the geometry features of the debris and the roof structure. However, recent studies have mainly focused on surface damage; little research has been conducted on structural damage. This paper presents an original 3D shape descriptor of individual roofs for detecting roofs with surface damage and roofs exhibiting structural damage by identifying spatial patterns of compact and regular contours for intact roofs, as well as jagged and irregular contours for damaged roofs. The 3D shape descriptor is extracted from building contours derived from airborne LiDAR point clouds. First, contour clusters are extracted from contours that are generated from a dense DSM of individual buildings derived from point clouds. Second, the shape chaos indexes of contour clusters are computed as the information entropy through a contour shape similarity measurement between two contours in a contour cluster. Finally, the 3D shape descriptor is calculated as the weighted sum of the shape chaos index of each contour cluster corresponding to an individual roof. Damaged roofs are detected solely using the 3D shape descriptor with the maximum entropy threshold. Experiments using post-event airborne LiDAR point clouds of the 2010 Haiti earthquake suggest that the proposed damaged roof detection technique using the proposed 3D shape descriptor can detect both roofs exhibiting surface damage and roofs exhibiting structural damage with a high accuracy.

A greyscale voxel model for airborne lidar data applied to building detection

“…(3) Clustering. This method first associates each lidar point with a feature vector, which consists of geometric and/or radiometric measures and then segments lidar points in feature spaces by a clustering technique such as k-means, maximum likelihood or fuzzy clustering (Filin, 2002;Hofmann, 2004;Vosselman et al, 2004;Filin and Pfeifer, 2006;Sun and Salvaggio, 2013;Kong et al, 2014;Zhao et al, 2014;Song et al, 2015;He et al, 2016;Kim et al, 2016;Cao et al, 2017). (4) Filtering.…”

Building Point Detection from Vehicle-Borne LiDAR Data Based on Voxel Group and Horizontal Hollow Analysis

“…[1]. The results of related studies have been widely used in fields such as intelligent building [2], virtual roaming [3], ancient architecture conservation [4], disaster management [5], and others. A laser scanning system has the unique advantage of providing data of directly measured 3D points in building extraction and 3D reconstruction [6].…”

Earth Observations for Geohazards: Present and Future Challenges

“…Using airborne LiDAR data, He et al [48] develop a 3D shape descriptor to detect surface-and structure-damaged roofs. Using airborne oblique images, Vetrivel et al [49] develop a Visual-Bag-of-Words (BoW) based damage classification to detect structure-damaged areas.…”