A novel approach for pavement texture characterisation using 2D-wavelet decomposition

In this paper, a 2D-wavelet approach is proposed to characterize the pavement texture acquired from the 3D laser scanner. The texture is decomposed into nine levels (from 0.1 mm to 25.6 mm) to extract the feature at multiscale. The Relative Energy (RE) and 2D-Entropy are calculated as indicators to represent the mixture surface texture distribution properties. Through conducting the 2D-wavelet decomposition on eight types of mixtures, it is found that the decomposition results perform well in recognizing the aggregates and predicting the gradation distribution. A wearing test for 100 h is also conducted and the results reveal that the mean profile depth (MPD) is highly correlative to the energy of the macro-texture while the RE and 2D-Entropy of the micro-texture decrease during the polishing process. Moreover, the mixture surface texture variates differently in the two directions: the RE of micro-texture deteriorate faster in the y-direction (the wheel movement direction) than the x-direction. The results show that the indicators can be used to measure the pavement texture variation due to traffic wear during service life, demonstrating that the novel 2D-wavelet approach the potential to evaluate road performances, such as skid and wear resistance.

Automated summary

A 2D-wavelet approach is proposed in this paper to characterize pavement texture, which can effectively recognize aggregates, predict gradation distribution, and evaluate road performances.