Automatic detection of delamination on tunnel lining surfaces from laser 3D point cloud data by 3D features and a support vector machine

Abstract

AbstractA completely automatic algorithm for accurately detecting delamination on tunnel concrete lining surfaces using laser 3D point cloud data is first proposed to facilitate tunnel lining inspection. A mobile mapping system (MMS), which mounts laser sensors and a positioning system, is utilized to measure the geometries of the surfaces at high speed. The algorithm consists of two steps: extraction of the 3D shapes of anomalies and discrimination of delamination from appendages by a support vector machine (SVM). The article focusses on the second step. On tunnel linings, there are many conspicuous appendages such as cables, lights, signs, and water guides which mask the features of delamination. In this study, straightness, a novel 3D feature, is introduced to realize accurate discrimination. An automatic algorithm based on the SVM is developed and validated using real tunnel data, showing an accurate delamination map.