Quantitative evaluation of peeling and delamination on infrastructure surfaces by laser signal and image processing of 3D point cloud data
Tsukasa Mizutani, Takahiro Yamaguchi, Tomoshi Kudo, Kazutomo Yamamoto, Tetsuya Ishida, Yoshifumi Nagata, Hinari Kawamura, Tomoaki Tokuno, Kiyoshi Suzuki, Yuuya Yamaguchi
Abstract
In this article, an algorithm for the automatic and quantitative evaluation of peeling and delamination on infrastructure surfaces from laser 3D point cloud data is proposed using state-of-the-art signal and image processing methodologies. A Mobile Mapping System (MMS) enables road administrators to collect the geometry of tunnel and road surfaces highspeed. The algorithm consists of two steps: 1D signal processing and 2D image processing. The peaks of anomalies are extracted by the envelope of a signal using a Hilbert transform. Anomalies detected in each direction are integrated into a map and smoothed by a morphology transform. The algorithm was validated by an experimental tunnel and road surface data. The proposed algorithm accurately and efficiently detects delamination and estimates the areas and depth of peeling and potholes by a real time analysis approach.