Method for dynamic segmentation of asphalt pavement performancebased on kernel density estimation(PDF)
长安大学学报（自然科学版）[ISSN:1006-6977/CN:61-1281/TN]
- Issue:
- 2020年2期
- Page:
- 10-20
- Research Field:
- 道路工程
- Publishing date:
Info
- Title:
- Method for dynamic segmentation of asphalt pavement performancebased on kernel density estimation
- Author(s):
- XU Zhepu; YANG Qun
- (Key Laboratory of Road and Traffic Engineering, Ministry of Education,Tongji University, Shanghai 201804, China)
- Keywords:
- road engineering; asphalt pavement; dynamic segmentation; kernel density estimation; pavement performance
- PACS:
- -
- DOI:
- -
- Abstract:
- Aimed at the problem that the currently used static road segmentation method cannot make full use of the basic detection data to guide the actual maintenance work, a new dynamic segmentation method for asphalt pavement based on kernel density estimation was proposed. Based on the fine pavement condition data acquired by rapid detection equipment, the pavement damage rate, international roughness index and rut depth data were analyzed by a line feature kernel density estimation method, so that a continuously changing pavement condition map with the kernel density value as the index could be obtained. On the basis of the existing pavement condition evaluation criteria, a new evaluation criteria for pavement condition with the kernel density value as the index was established, thereby achieves the dynamic segmentation and evaluation for the asphalt pavement. Through the reclassification and combination method, the three indices of pavement damage, roughness and rut depth based on the kernel density value could be combined，and to further obtain a comprehensive segmentation scheme show multiple indices simultaneously. South Heping Road in Jinjiang was taken as an example, the application of this method in practice was demonstrated. At last, the proposed method was compared with the static road segmentation method and a dynamic road segmentation method based on cumulative difference approach. The results show that the proposed method can take the spatial autocorrelation of road condition data into consideration and obtain a fine segmentation result, which segments the road into sections, making the road conditions inside the same road section as similar as possible, while the ones between neighbor sections as different as possible. The pavement condition can be evaluated from different point of views, and from the view of the three indices individually or of the comprehensive one, which shows the overall distribution of the pavement condition and the urgency level of maintenance in different sections, and meanwhile retains the details of the basic detection data. The research also shows that this method can link up with the existing pavement management system well, which is conducive to the promotion and application of the dynamic segmentation method. 8 tabs, 6 figs, 29 refs.
Last Update: 2020-04-09