Real-time Distress Screening in Automatic Pavement Condition Data Collection

Pavement condition data collection is one important part for best cost-benefit pavement management and maintenance in all transportation agencies. Large amount of pavement distress data of crack, rutting, potholes, etc. are surveyed to better manage and maintain the pavement asset. However, most of the data collected have no distress inside at all. It is a waste of resource to sample and process it for on-line data acquisition and off-line post-processing. The quick screening of being crack or not could adaptively improve the imaging interval and exposure of acquisition, and relieve the amount of data for quality review. This paper proposes the real-time screening of pavement crack to demonstrate the feasibility of avoiding the irrelevant data during pavement condition survey. The proposed algorithm takes advantage of and integrates the accuracy of multi-scale crack integral profiling (M-CIP) and high speed of crack progressive screening (CPS). Based on the Gaussian-smoothed profile, crack integral profiles (CIPs) are formulated to characterize the transverse continuity of crack pixels. CIP value at a pixel reflects the degree of being crack of a certain width. It is calculated quickly using the innovative integral profile. M-CIP is then conducted to get the optimal crack characterization among different scale widths using the running mean. A fast crack progressive screening strategy is finally proposed to mimic the manual drawing of significant crack indications by the connected M-CIP analysis and crack connectivity check on a profile-by-profile basis. The proposed algorithm is tested on a diverse set of actual pavement images taken on interstate highway G-4 near Beijing provided by the National Center of Measurement for Equipment of Roads and Bridges, Research Institute of Highway of China under varying lighting conditions. Experimental results show the proposed algorithm runs 62 nanoseconds per pixel and 97ms for a 1023*1528 image on a normal personal computer with zero false negative (FNs). It is fast and accurate enough for real-time and reliable crack screening at an acquisition speed of 100 Km/h. The proposed algorithm is promising for practically implementing distress screening of various types in pavement condition data collection.

• 
• Supplemental Notes:
  • This paper was sponsored by TRB committee AFD20 Pavement Monitoring and Evaluation.
• Corporate Authors:

  Transportation Research Board

  500 Fifth Street, NW
  Washington, DC  United States  20001
• Authors:
  • Jing, Genqiang
  • Huang, Yuchun
  • Yin, Wenbin
  • Wang, Qinglei
• Conference:
  • Transportation Research Board 94th Annual Meeting
  • Location: Washington DC, United States
  • Date: 2015-1-11 to 2015-1-15
• Date: 2015

Language

• English

Media Info

• Media Type: Digital/other
• Features: Appendices; Figures; References; Tables;
• Pagination: 16p
• Monograph Title: TRB 94th Annual Meeting Compendium of Papers

Subject/Index Terms

• TRT Terms: Algorithms; Automatic data collection systems; Pavement distress; Pavement management systems; Real time information
• Geographic Terms: Beijing (China)
• Subject Areas: Data and Information Technology; Highways; Maintenance and Preservation; Pavements; I60: Maintenance;

Filing Info

• Accession Number: 01551930
• Record Type: Publication
• Report/Paper Numbers: 15-2934
• Files: PRP, TRIS, TRB, ATRI
• Created Date: Jan 29 2015 9:17AM