Development of a Repair Estimation Calculation System for Determining Damage Evaluation of Residential Buildings Using AI

Abstract

The amount of survey work required for post-earthquake residential damage assessment is enormous, and the degree of damage is assessed by administrative staff based on visual inspection of the house's exterior. Therefore, this study aims to automate the residential damage assessment survey using AI to improve efficiency. For this purpose, this paper attempts to construct a system to calculate the repair estimates for damage rate calculation. Specifically, we developed a program that calculates repair estimates based on the classification of building exteriors based on area suggestions. Experiments were conducted on the developed calculation program, and we examined the predictability of the results of the judgment of the repair cut-off.