Road condition assessment: A framework for automatic detection of surface flaws

Abstract

Road abnormalities such as cracks, unevenness, potholes, and manholes are increasing the number of road disasters in today’s world, particularly in nations like India. Accidents and irreplaceable loss result from uneven and damaged roadways, as well as unneeded openings. The introduction of more Big Data sources through citizen recording devices has created a new foundation for public infrastructure management and control, as well as policy design. Roads that are maintained on a regular basis are less likely to be involved in accidents. However, manually inspecting road damage is costly, time-intensive, and requires a large amount of manpower. Automatically detecting and reporting the presence of potholes, manholes, and other anomalies such as cracks to the appropriate departments can aid in the recovery of road conditions. Detailed real-time performance object detection frameworks (YOLOv5 and RCNN) for detections of potholes are presented. The main objective of this manuscript is to propose a framework that utilizes machine learning and deep learning models for detecting surface flaws.