Abstract
This study is to develop a detector that automatically detects deteriorations such as cracks, efflorescence, and chalk markings from the photographs of concrete structures, using convolution neural network which is a kind of deep learning. Firstly, photographs of concrete were collected for the learning data. Secondly, images of single cracked part, multiple cracked part, efflorescence part, efflorescence with cracks part, chalk marking part, construction joint part, surface part, rust part and others part were producted from these photographs for the learning dataset. Thirdly, a classifier to classify images into these 9 class was developed using the learning dataset and convolution neural network. Fourthly, the automatic detector was constructed using this classifier. Finally, a web system to receive a photograph and return detections was implemented based on Twitter.
