Research on Method for Evaluating Displaced Joints of Sewer Pipes Using Deep Learning

In Japan, the number of sewer pipes with a useful life 50 years is expected to increase in future, and efficient methods are required maintain them. Various abnormalities occur pipes. particular, displaced joints must be inspected repaired because they major factors resulting road cave-ins. current inspection investigation pipes, surveyors operate closed-circuit television (CCTV) camera while checking monitor on ground record presence degree abnormality field or office. However, this method has certain problems that it time consuming expensive visually evaluate variation judgment results among surveyors. research, we propose for detecting evaluating from images using deep learning address these issues. Then, experiments will conducted clarify its usefulness.