The article discusses the methodology for recognizing road surface defects for clustering and ranking road sections. Relevance and purpose. The relevance is due to the fact that in order to improve road safety and plan road work, it is necessary to assess road surface defects according to the degree of danger for vehicles. The goal is to warn drivers and synthesize routes for road maintenance services during the planning of repair and restoration work. Materials and methods. The methodology includes methods, models and algorithms for recognizing road surface defects, clustering and ranking road sections by operational condition to assess the risks of road incidents. Results. To achieve the goal of the research, the following tasks were solved: a) processing images with defects, b) segmentation and recognition of defects in images using the IoU-HOG-ACM-BoVW algorithms and the MaskR-CNN neural network, c) spatial clustering of areas with similar operational conditions, d) ranking of clusters and road sections according to the number and degree of risk of incidents, percentage of defects, e) synthesis into the selection of routes for pavement repair using the ant colony algorithm and the hierarchy analysis method. Conclusions. The methodology is implemented in an intelligent system for monitoring road transport infrastructure. The main advantage is the creation of an integrated approach for solving problems of recognition, clustering, risk assessment and ranking of road surface defects with the subsequent synthesis of optimal routes in the process of planning work to restore the road surface.