Pavement damage detection and evaluation based on UAV image and improved AHP

As a crucial component of road infrastructure, the pavement significantly impacts the comfort of vehicle and pedestrian traffic. Various factors such as traffic flow, weather conditions, and material quality can lead to pavement damage. This paper utilizes UAVs to collect road damage images and establishes a classification and evaluation model by using image recognition algorithms and the analytic hierarchy process (AHP) theory. To verify the effectiveness of the model, a road in Chongqing was selected as the experimental object. This model can precisely and effectively distinguish different types of pavement damage, such as ruts, grooves, and so on. The results show that the classification accuracy of the model is 92.5%, which has good detection performance compared with other models. In the research, it was found that the contrast of rut damage images varies greatly, ranging from 0.08 to 0.32. This difference may be due to the larger damage area present in the rut image compared to the other images. Considering groove and crack features are similar, there is a 66.7% chance of identifying one as the other.