Applying Computer Vision Technology to Triaxial Deformation Monitoring of Remedial Construction for a Landslide

Background & Objective: The paper explores a new instrument of computer vision to measure three-dimension deformation with an Internet of Things (IoT) system including Raspberry Pi, digital cameras and OpenCV programs in laboratory and field testing so as to monitor the potential deformation of a structure drainage well in a landslide. Methods: A chessboard pattern is detected in the image by the camera so that pixels of chessboard cornors can be recognized by OpenCV programs. X-direction, Y-direction and Z-distance changes can be casulated by the similar triangles relationship of camera pixels. For laboratory testing, standard deviations of the measurement were approximately 0.01 cm. Results: For field testing, the study installed four sets of Raspberry Pi in a drainage well within a landslide and employed OpenCV programs to interpret pixel changes of chessboards at four levels of the draiage well. Conclusion: Overall, the instrument can be employed for triaxial deformation monitoring of the construction in the field effectively and automatically.