Research On Pavement Distress Detection And Classification Using Wavelet Analysis

The growth in the amount of vehicle miles traveled (VMT) of highway cause the maintenance to be heavily increased. Traditionally, pavement condition surveys are artificial surveys. Apart from being labor-intensive, time-consuming, low efficiency, the method is also quite subjective, and the result is produced slowly, the explanation needs a long period, which can not meet the modern highway conservation requirements. In recent years, to overcome the above limitations of the subjective visual evaluation process and meet the modern, large-scale, high speed and quality of road maintenance management requirements, several attempts have been made to develop more automated pavement inspection systems in the pavement image processing. Thus pavement distress image preprocessing, distress detection, identification, evaluation and classification become a hot research field.In this paper, we completed the pavement automatic detection, classification and disease assessment by the digital image processing technology and the specific contents are as follows:(1) The cracks in the asphalt pavement for the image preprocessing algorithms were studied, including image enhancement and image segmentation.(2) Distress detection algorithm and classification algorithm based on the Radon transform were studied.(3) The pavement distress evaluation algorithm based on wavelet analysis was studied.Project Team cooperate with the Jiangxi Provincial Quality Supervision Station of the Highway Administration Bureau, so we have accumulated a large number of pavement images of Jiangxi highway. The experimental results show that, the automated pavement inspection system which we studied can detect the crack image exactly, and the correct rate is about 97%, and the result of the classification is also idea, only few errors between block cracks and alligator cracks.