3d Model Reconstruction Using Aerial Photos and Application To Detection of Buildings Damaged Due To Earthquakes

Abstract eng:
This paper present a disaster monitoring methodology based on the combination of Structure from Motion (SfM) and Multi View Stereo (MVS) techniques; the SfM-MVS technique is one of the three dimensional shape reconstruction technology based on the assembling of photos acquired from different sources. In order to study the potentiality of the technique for damage detection, first the SfM-MVS accuracy has been investigated. For such purpose, the dataset over the damaged areas in Hiroshima by a large-scale debris flow occurred on August 20th, 2014 was used. A number of photos were taken by an aerial vehicle just after the event, and LiDAR measurement was also carried out at the same time. The aerial photos were used for the creation of the surface model by SfM-MVS, and LiDAR data were used for the validation of the height values. For examining the height accuracy, we applied SfM to the dataset and observed the numbers of feature points extracted from input image. Then, we created point cloud by MVS and calculated the difference of vertical values and its error from the created model and actual height information from LiDAR data. Next, for the application to earthquake damage detection, we applied SfM-MVS to the affected areas by the 1995 Kobe earthquake and evaluated the extraction of severely damaged buildings.

• Export as: BibTeX | MARC | MARCXML | DC | EndNote | NLM | RefWorks
• Add to your basket