Road damage happens everywhere, such as road surface cracks and missing lane markings, which impacts traffic efficiency and sometimes even leads to dangerous situations. Every year, road authorities spend plenty of resources to detect road damage. However, due to human error that arises from manual work, results are not always optimal. To address this concern, a growing number of organizations use computer vision technology to improve detection quality and efficiency. By applying AI algorithm to field-collected videos and images, the road defects can be efficiently detected and classified. This makes the inspection process automated and provides evidence of predictive maintenance to road authorities.