An Automated Filter Bank-Based Pavement Crack Detection System Incorporating Standard Compression Coders

• Published in: International Journal of Pavement Research and Technology Vol. 5; no. 3; pp. 176 - 186
• Main Authors: Lettsome, Clyde A, Tsai, Yichang (James)
• Format: Journal Article
• Language: English
• Published: Jhongli City 中華鋪面工程學會 01.05.2012; Springer Nature B.V
• Subjects: Automated; Automation; Coders; Compressing; Concrete pavements; Cracks; Filters; Flaw detection; Pavements; Segmentation; Sensors; Stress analysis; Wavelet; JPEG2000; Pavement crack detection; Filter bank; Low bit rate; Pavement distress segmentation; Said Pearlman SPIHT Coder
• ISSN: 1997-1400; 1996-6814; 1997-1400
• DOI: 10.6135/ijprt.org.tw/2012.5(3).176

Wavelets/filter banks have become a popular area of research, especially in pavement distress, because they can be used concurrently to separate the subbands of signal frequencies and analyze the data in time/spatial domains. Because of this, filter banks have been studied in automated pavement distress detection and distress segmentation systems research. However, incorporating standard compression coders like Set Partitioning in Hierarchical Tree (SPIHT) or Joint Photographic Experts Group 2000 (JPEG2000) into automated filter bank-based distress detection and segmentation systems has not been presented. An automated filter bank-based pavement crack detection system that can effectively incorporate Standard Compression Coders into a filter bank-based automated crack detection and segmentation system has been proposed in this paper. This is the first known attempt to incorporate standard coders into a system. The proposed system has been validated using real pavement images provided by the Georgia Department of Transportation (GDOT). Preliminary results show that the proposed system can provide usable segmentation results, even at a low bit rate. This allows researchers to reduce data storage requirements and, most importantly, increase the speed over the Embedded Zero-tree Wavelet (EZW) used in previous filter bank-based automated pavement distress systems. Future research is also discussed. [PUBLICATION ABSTRACT]