EVALUATION OF EFFECTS OF TRAINING DATA FOR DEEP LEARNING-BASED DETECTION OF PELLING AND REBAR EXPOSURE

Abstract

A bridge inspection needs a lot of costs. It causes a lack of engineers and budget. So some local governments couldn't complete the bridge's aggressive preventive maintenance in Japan. Recently, deep learningbased damage detection methods have been studied by many researchers. This kind of method could detect the damage of a bridge by photo image. On the other hand, the effectiveness of the quality of training data is not discussed.

In this study, we targeted the automatic detection process of peeling and rebar exposure in bridges. We evaluated the effects of the ratio of the negative example data in a training dataset. We trained to detect models with different non-annotated data ratios. As a result of comparing the detection results of each model, the influence of the negative example data on the training data was confirmed.