Critical Assessment of Detecting Asphalt Pavement Cracks under Different Lighting and Low Intensity Contrast Conditions Using Emerging 3D Laser Technology

Abstract

After decades of research and development, a fully automated system for pavement crack detection with an intensity-based two-dimensional (2D) imaging data acquisition system under different lighting and low intensity contrast conditions still remains a challenge. With the advances of sensor technology, a three-dimensional (3D) laser technology that can collect high-resolution 3D continuous transverse pavement profiles for detecting cracks on the basis of their 3D elevation rather than 2D intensity has become available. This study, sponsored by the U.S. Department of Transportation (US DOT) Research Innovative Technology Administration (RITA) program, evaluates the feasibility of using emerging 3D laser technology to detect cracks under different lighting and poor intensity contrast conditions. For this purpose, 3D surface data from laboratory tests on cracks ranging from 1 to 5mm wide and tests on the actual pavement cracks of different widths and depths and under different lighting conditions were performed. The dynamic-optimization-based crack segmentation method was employed to detect cracks. A linear-buffered Hausdorff scoring method was used to quantitatively evaluate the crack segmentation performance by comparing the detected cracks with the manually established ground truth. The preliminary test results show that: (1) the 3D laser system can effectively detect cracks equal to and greater than 2 mm wide under controlled laboratory environment, (2) it can achieve consistent results when detecting cracks under different lighting conditions (i.e., nighttime, daytime with shadow, and daytime no shadow), and (3) it can detect cracks with low intensity contrast. Future research is recommended.