Applying Computer Vision Technology to Health Monitoring for Underground Structure in a Landslide Area

Abstract

The study employs computer vision technology to explore an innovative instrument for deformation monitoring with an Internet of Things (IoT) system composited of Raspberry Pi, digital cameras and chessboards. Through laboratory and field testing, the instrument was used for settlement measuring for an underground structure in a landslide area in order for more cost-effective methods of slope monitoring. For laboratory testing, the resolution of the instrument was approximately 0.01 cm. For field testing, a set of the instrument was installed in a drainage gallery that was constructed at an 80-m-depth location within the landslide area. The accuracy of field instruments can reach to 0.08 cm when the optimal distance between the camera and the chessboard was 20 m. Moreover, the instrument costs only one tenth prices of traditional monitoring devices such as electrical tilt meters and extensometers. Overall, the instrument based on computer vision technology can be automatically and efficiently employed for settlement monitoring in the landslide with the IoT system.