Automatic detection of asphalt pavement thickness: A method combining GPR images and improved Canny algorithm

• Asphalt pavement thickness is detected by GPR system for its advantage of non-destructive. • The traditional Canny algorithm, the connected region detection algorithm and the improved Canny algorithm were compared by simulated and actual images. • The improved Canny algorithm was used to detect GPR images, which has a smaller error compared with the traditional Canny algorithm and the connected region detection algorithm. The traditional drill core sampling method used for pavement thickness detection is increasingly difficult to meet the increasing demand for pavement detection. At the same time, ground penetrating radar (GPR) have shown superiority in pavement non-destructive detection for its fast detection, safety and high efficiency. In this study, the traditional Canny algorithm was improved by combining wavelet denoising, intercept method and artificial bee colony algorithm, and the improved Canny algorithm was compared with the traditional Canny algorithm and connected region detection algorithm by combining simulated images and actual images. The detection results showed that the improved Canny algorithm had better performance, and the relative error was about 3.82%, which can realize the fast and intelligent detection of asphalt pavement thickness.