Abstract: Aimed at the problems of low accuracy of pavement crack classification and high cost of pavement disease location in road detection , a pavement disease recognition and location method based on monocular vision is proposed. The method was implemented in two stages. In the first stage, YOLOv5 was used to complete the detection and output the detection frame and initial category. In the second stage, the category adjustment and monocular positioning algorithm were used to convert the road picture to the orthographic perspective , calculate the location and size of the disease , and adjust the initial category. Experiments show that the classification accuracy of the algorithm can reach 90% , the positioning and measurement accuracy of the target can reach centimeter level , and the computational power requirement is low. Only a single camera is used to detect and locate pavement diseases at the same time , which can meet the requirements of real-time detection of pavement maintenance engineering.