Contribution of Dual-tree Complex Wavelet to Three-dimensional Analysis of Pavement Surfaces

Pavement industries in Japan enhance the application of three-dimensional (3D) modeling as a part of the policy called "i-Construction" which applies 3D point cloud data to pavement works. However, distinctive approaches in terms of signal processing are required in practical applications of 3D point clouds to pavement surfaces. An effective and efficient filtering algorithm is then necessary for the analysis. The purpose of this study is to verify the ability of the dual-tree complex wavelet transform (DTCWT) applied to the 3D point clouds measured for pavement surfaces. Unlike conventional continuous and discrete wavelet transforms, the DTCWT allows nearly shift invariant and directional de-composition in two and higher dimensions with less redundant manners. This study conducts a field experiment at a test site paved with precast concrete blocks for the verification. The result shows that the DTCWT provides superior decomposition algorithm to the discrete wavelet transform by enabling effective filtering based on the directional multiresolution analysis. Finally, the performance of DTCWT is proved for the identification of pavement distress and deformation effectively and reasonably in terms of wavelengths and locations in this study.