In Japan, about 70% of the land is covered with forests, and many urban areas tend to be located along the sea. Therefore, road networks connecting urban areas often pass through mountainous areas. Road networks in mountainous areas include cut slopes, and many shotcretes have been adopted as slope protection works at that time. Many shotcretes were constructed during the period of high economic growth, and they are aging all at once. In the maintenance of shotcrete, it is important to grasp the deformation and take appropriate measures. Deformations of interest on the shotcrete include cracks, peeling, and floating. Of these deformations, it is difficult to visually confirm the void, and it has been confirmed by a hammering test. However, due to the shortage of inspection engineers and the financial difficulties of the national and local governments, there is a limit to the continuous diagnosis of the void by hammering tests. In this study, we developed a deep learning model using infrared images acquired from an infrared camera mounted on a UAV, taking advantage of the difference in heat capacity between the void part and the not void part. In this study, we adopted CNN (Convolutional Neural Network), which is one of the deep learning methods and is used in image recognition, and learned the model by featuring the temperature distribution around the void part. In addition, as a result of verifying the deep learning model by core sampling, it was confirmed that damage can be detected with high accuracy.

mortar sprayed slope;UAV;Deep learning