Research On Feature Extraction And Paramater Calculation Of Pavement Rutting

During the recent years, traffic transportation especially the highway construction has achieved rapidly development in our country. As increasing of the mileage of highway and the amount of heavy vehicles, it has become an extrusive problem in pavement transverse profile distortion. Thereinto, the analysis of rutting distress becomes an important composition for pavement performance assessment and a fundamental element for maintenance decision.At present, the rutting measurement adopted manual operation in our country. This method is simple and reliable, but with low efficient or the obtained data including much man-made factor. With the development and application of the non-contacted technology, such as laser and ultrasonic technology, the velocity of rutting measurement has greatly raised with automatic inspected technology, which based on the highness measurement of accidented road surface. The curve which only using some scatter dots can not reflect the real road transverse profile due to the limits of sampling, thus it can not provide enough precise result by this way. Thanks to the maturation of machine vision technology, it performed many advantages and found its position in pavement inspection, which made rutting measurement in a precise, effective, real time and non-stop way. But the results mixed with other pavement transverse feature signals except rutting owing to the scanning way. It becomes a key topic that how to extract rutting feature from pavement transverse profile signal and its parameter calculation. Little research about this aspect has been done in and outside our country. Aimed at the above problem, the below tasks has been done in this article.(1) The theory of the 3-D pavement measurement by structured light which based on the technology of Laser trigonometry been analyzed. Data acquired system structural composition and its working principle also been introduced in this article.(2) Based on the analysis of transverse profile signal characteristics and the performance of different filters, the algorithm combing the wavelet with median filter was put forward to extract rutting feature in this article. The procedure of proposed algorithm is performed to eliminate the balanced stochastic noises as well as other feature mixed in road transverse profile, to make sure the rutting feature extraction un-distorted and available.(3) Based on the testing rules of rutting local measurement, and followed by manual rutting measurement in which need to find the top datum line of the road profile, the algorithms of looking for suppositional ruler and wire based on straight-edge model and wire model was put forward. And the study of rutting parameter calculation, such as rutting depth, width and location, has been carried out.The results showed that the proposed methods are efficient, simple and available in rutting extraction and parameter calculation. The results can provide information to perform assessment and maintenance with important signification for road administration.