Estimation of Road Damage for Earthquake Evacuation Guidance System Linked with SNS

EasyChair Preprint 4043

Abstract

The damage caused by earthquakes is very serious in Japan. One of the most common causes of earthquake damage is delayed escape due to obstructed passage. Rapid evacuation is very important during an earthquake. However, the pathway may be obstructed by obstacles during an earthquake. In this study, we propose an evacuation guidance system to evacuate users from a disaster area by using information from SNS and other sources. In this study, we propose an evacuation guidance system to evacuate users from a disaster area by using information from SNS and other sources. The proposed system enables users to complete their evacuation in the shortest and safest route. However, it is necessary to estimate the damage of the corridor to realize the system. Thesis In order to reduce the computational cost, transfer learning was performed using VGG16. As a result, the classification was achieved with an accuracy of 78%. This method can be used to automatically identify the damage in the corridor and help in determining the evacuation route.

Keyphrases: Earthquake, Evacuation guidance system, Support Vector Machine

@booklet{EasyChair:4043,
  author    = {Yujiro Mihara and Rin Hirakawa and Hideaki Kawano and Kenichi Nakashi and Yoshihisa Nakatoh},
  title     = {Estimation of Road Damage for Earthquake Evacuation Guidance System Linked with SNS},
  howpublished = {EasyChair Preprint 4043},
  year      = {EasyChair, 2020}}