Dual-Light Inspection Method for Automatic Pavement Surveys

Abstract

Conducting a pavement survey is a time-consuming but necessary task to ensure the serviceability of road pavements. Many investigators have used image-processing methods to automate the survey process and to enhance the quality and accuracy of survey results. However, image-processing methods often mistakenly treat oil spillages, shadows, and road markings as distresses because their features are similar to those of distresses. Therefore, in this research, the authors proposed a dual-light inspection (DLI) method to reduce false alarms. The DLI involves four major steps: (1) image capture—a pair of images is retrieved from identical positions and orientations, but with different light setups; (2) image subtraction—the two images are subtracted pixel by pixel to obtain a subtracted image that represents the differences between the paring images; (3) image enhancement—an edge detection method is applied to retrieve the distress features; and (4) image classification—a classification algorithm is finally used to discriminate between images that include distresses and ones that do not. A field test was conducted to verify the DLI method. A total of 212 pairs of images were captured during nighttime, including images of alligator cracks (42 pairs), manholes (42 pairs), longitudinal cracks (58 pairs), spillages (34 pairs), and road markings (52 pairs). Twenty percent of the images (i.e., 45 pairs) were used as training sets to train the classification model. The remaining images were then used to test the accuracy of the classification model. The accuracy of the DLI method, which uses dual-light image pairs, was compared with that of the traditional method, which uses individual images. The DLI can significantly improve the accuracy in determining spillage (traditional: 18%, DLI: 82%) and road markings (traditional: 8%, DLI: 96%). The DLI is also reasonably accurate in determining other distresses including alligator cracks (traditional: 95%, DLI: 90%), manholes (traditional: 97%, DLI: 100%), and longitudinal cracks (traditional: 62%, DLI: 69%). These results indicate that DLI can become a reliable method for automatic pavement inspection.