Automatic Crack Monitoring System

Details

• Creators:
  Liu, Richard ; Chen, Xuemin ; Wu, Min
• Corporate Creators:
  University of Houston. Department of Electrical and Computer Engineering
• Corporate Contributors:
  Texas Department of Transportation. Research and Technology Implementation Office
• Subject/TRT Terms:
  Algorithms Digital Cameras Image Processing Monitoring Pavement Cracking User Interfaces (Computer Science)
• Publication/ Report Number:
  TxDOT 7-3997
• Resource Type:
  Tech Report
• Geographical Coverage:
  United States ; Texas
• Edition:
  Final Report
• Corporate Publisher:
  Texas. Department of Transportation
• Abstract:
  This report discusses the design and implementation of a Highway Crack Monitoring System (HCMS) used in a Texas multi-purpose load simulation (TxMLS) test site. The HCMS is a black-and-white CCD camera-based image-processing device that processes the image of the cracked pavements, extracts crack information and characterizes the cracks in terms of crack length and width. In this project, both hardware and software are developed. The hardware system is fully automatic with computer-controlled x-y motor motion devices and Hall-effect sensors. Many lab tests were conducted. Processed results show that an accuracy of more than 90% is achieved for crack width and length classification. The crack-processing algorithm developed in this project uses the method of feature extraction and object characterization. A user-friendly, Windows-based user interface is developed, which allows user to control the camera motion, calibrate the system, acquire crack images, process images, and store data. The accuracy of the object characterization largely depends on the reliability of the features extracted from the original image data. The HCMS system implements recurring thresholding as the basic segmentation algorithm, which adaptively determines the gray level threshold through an estimation-verification process. Distinct block operation is applied to improve the performance on the non-uniform property of the road image. Connected-component object recognition and other criteria are also implemented to distinguish the crack objects and remove the noise. The crack object in the binary image is characterized by object boundary processing. The perimeter of the object can be obtained by accumulating the moves of the pixels on the border contour. Border convolution is implemented to classify the border pixels and to determine the corresponding move length. With the area of the object, the length and width information can be obtained. The limitation of the algorithm is also discussed.
• Format:
  PDF
• Collection(s):
  US Transportation Collection
• Main Document Checksum:
  urn:sha-512:be26c184cde20435217d41878b7ffbf3edf65da446b33ca1637d7415facd6f0afda57304ce78c60e8d42df5098fbac0007f4806fa139d0d1753369ba74270384
• Download URL:
  https://rosap.ntl.bts.gov/view/dot/87080/dot_87080_DS1.pdf
• File Type:
  [PDF - 2.49 MB ]