Asphalt Pavement Crack Detection Using Image-to-Image Translation

Pavement plays a crucial role in transportation because it is a permanent surface for use in road networks. The health of the pavement ensures the safety and convenience of drivers and passengers. In the past few decades, pavement management systems have encountered challenges which often have produced solutions with excessive demand for resources, but low-accuracy results. New approaches must be developed in order to quickly and economically identify pavement failure, especially cracks. This paper proposes a fast and accurate method for segmentation of all types of cracks in asphalt pavement images based on generative adversarial networks (GANs). The proposed model learns the mapping between two domains of pavement images and images of segmented cracks. This approach does not necessitate any preprocessing or post-processing tasks, and the model generates new images without the need to classify each pixel. It detects cracks with high accuracy using a conditional image-to-image translation. In this study, the model took an average of 0.29 s to identify the cracks in each image. This outstanding crack identification had a precision of 85.76%, a recall of 89.81% and an F1-score of 87.72%.