Automatic Classification Of Pavement Based On Laser Data

With the rapid development of the transportation industry, the road maintenance management has been attracting more and more attentions. One of the most important things of the road maintenance management is road distress classification. The traditional method of obtaining pavement information already can not adapt to the development of the highways. Development of automatic identification system seems more important and urgent.This paper aim at the road distress classification, some methods are proposed in order to realize the classification. It firstly introduces the situation of the pavement damage automatic identification system both at home and abroad, and the pavement damage types. Based on this, some new technologies are proposed. And then laser detection technology is discussed. At last, this paper discusses the support vector machine for pavement damage of classification, and proposed the laser data classification based on the support vector machine (SVM).How to realize the pavement classification is the core of the pavement automatic identification system. In this paper, laser data feature extraction method and support vector machine classifier design are discussed.In the feature extraction, according to the laser data characteristic of the different pavement damages, this paper proposes the laser data density feature extraction and the damaged depth feature extraction method to describe the distress images. Compared with the traditional methods, the results confirmed that the new feature extraction method has better adaptability and recognition effect.In the design of classifier, this paper analyzes the advantage of the support vector machine for pavement damage classification, using the laser data of SVM classification method realize the classification and recognition of the pavement damage. Experiments show that, the laser data algorithm and SVM method is able to further improve the accuracy.According to the above, this paper studies the pavement damage classification. The results of the study indicate that: the laser data can fully manifested the concrete pavement damage information and closely related with the pavement damage classification. The laser data algorithm and SVM method not only be able to identify holes, bumps etc, but also can effectively enhance the accuracy of classification.