Tunnel Lining Crack Detection Method Based on Android Mobile Phone

To solve the problems of inefficiency, manual dependence, and subjectivity of traditional tunnel lining crack detection techniques, as well as the limitations of high cost and difficulty in popularization of high-end technologies such as ultrasonic and radar detection, this paper proposes a tunnel lining crack detection method based on Android cell phones. Deep learning techniques are utilized to obtain the optimal weight file by training the YOLOv5s detection model, and subsequently, the weight file is converted to NCNN format for Android devices. Application software was developed with the help of the Android Studio tool to realize the deployment and application of the tunnel lining crack detection model on Android phones. The inspector takes an image through the cell phone’s camera function and selects the crack image from the cell phone album to start the crack recognition function based on the CPU or GPU driver with one click, which provides an efficient and economical solution for tunnel maintenance and safety management.