RESEARCH ON METHOD FOR VISUALIZATION OF CRACKS ON SABO FACILITIES USING UAV

Abstract

In Japan, in order to maximize the effectiveness of sabo facilities, periodic inspections are conducted. However, visual inspections involve a great deal of labor and risk to check for cracks and other damage and deterioration. Against this background, the Ministry of Land, Infrastructure, Transport and Tourism (MLIT) is aiming to establish a safe and efficient inspection method using UAV in order to reduce the burden on inspectors. However, a method for automatically detecting cracks in images of sabo facilities taken by UAV has not yet been established. In this research, we propose a method to visualize cracks by using a heat map based on a learning model constructed from existing image classification models and data sets. We also evaluate the usefulness of the proposed method through experiments.