Evaluation of Three‐Dimensional Shape Signatures for Automated Assessment of Post‐Earthquake Building Damage

This paper proposes two new measures of three‐dimensional shape signatures based on the slope and aspect of three‐dimensional triangles, and evaluates the performance of three‐dimensional shape signatures derived from distance, area, angle, volume, slope, and aspect for assessing post‐earthquake building damage using simulated building models with flat, pent, gable and hip roofs. Three scenarios of post‐earthquake building damage are tested using simulated light detection and ranging (LiDAR) points generated on the building roofs. Dissimilarity, sensitivity, and computational cost of the three‐dimensional shape signatures are also discussed. The results show that a combination of three‐dimensional shape signatures derived from slope and aspect can provide better detection of post‐earthquake building damage compared with those derived from distance, area, angle, and volume. The paper demonstrates a promising method for rapid assessment of post‐earthquake building damage using existing three‐dimensional urban models and post‐earthquake LiDAR data.