Automatic Recognition And Intelligent Evaluation For Expressways Pavement Distress

Transportation is the basis for the development of social economy and is also a powerful motive force of economic development and social improvement in China. 80s of last century, highways are achieved a breakthrough, and are entered the high-speed construction phase quickly in China. In the past 11 years, China's highways are increased from 652 kilometers to nearly 30,000 kilometers during 1992 to 2003. To the end of 2009, the mileage of China's highway has reached 65,000 kilometers. The total traffic mileage has become second in the world only next to the United States. With the development of economic society, some new requirements of scientific road maintenance and from corrective to preventive road maintenance have been proposed by people. Therefore, it has become an urgent problem to solve that pavement performance was assessed objectively and accurately.The study is carried out from the following four aspects: pavement image preprocessing, pavement image segmentation, pavement performance evaluation and pavement management system. The source of pavement surface distress image noise is analyzed firstly, and noise has been classified. This study points out that pavement surface distress noise is mainly divided into impulse noise and Gaussian Noise. And then,based on the noise detection of image, the traditional noise filtering method was improved and Pre-process Algorithm of pavement surface distress image based on noise detection is brought forward. This algorithm is better than other algorithms at filtering noise and preserving the detail of image.After analyzing the gray distribution features of pavement surface distress image, considering the general approach of image segmentation of pavement surface distress image, a new image segmentation algorithm of pavement surface distress image based on Generalized Structuring Element is proposed. Based on the combination of dynamic grade threshold Image Segmentation and grey value threshold Image Segmentation, and used as pavement image's segmentation threshold, experiments prove that the proposed algorithm is robust for pavement image background complication, shadow, illumination variance and different contrast level. Finally, the algorithmic feasibility is proved by computer simulation.Comparing the virtues and defects of Regression Analysis, Fuzzy Multiple Judgment, a new pavement performance evaluation algorithm based on Entropy Weight Radar Chart Theory is proposed. The algorithm has good performance and the result of simulation is consonant with real pavement.Finally, a digital Pavement Management System based on GIS is built. After detailedly analyzing overall frame work and designing every function module in the system, this system is developed on the platform of ArcGIS Server9.3, using C# development language and Oracle10g spatial database.