Pavement Distress Analysis based on Dual-Tree Complex Wavelet Transform

Abstract

Pavement distress identification and classification are critically important for pavement health management. This paper presents a new method for pavement distress identification based on the dual-tree complex wavelet transform (DT-CWT). It takes a multi-scale and multi-resolution approach to decomposing a pavement image into multi-level subbands, with high frequency subbands containing distress features selected as the subbands of interest. After thresholding is performed on wavelet coefficients to reduce noise, multi-level sub-images were constructed through inverse DT-CWT. The merit of the presented method rests on the shift invariance and good directional selectivity in DT-CWT while maintaining high computational efficiency. Numerical and experimental analysis results confirmed its substantial performance enhancement over ordinary Discrete Wavelet Transform (DWT), as commonly reported in literature.