Pothole detection and cost estimation - Research Paper

Abstract In this paper, An advanced pothole detection method is proposed in this project with accurate calculation of the dimensions of the potholes, estimation of their volume, and determination of the filling cost of these potholes by using video inputs from road surfaces. Our system uses model object detection models to process every video frame to precisely identify potholes, and then me the dimensions as concerning length, breadth, and depth by using depth estimation model. Specifically, the system applies a technology known as monocular depth estimation, which allows for the estimation of the precise depth of every pothole directly from the feed obtained from one camera, enabling the calculation of volume to be repaired. Calculating the volume of each pothole, it computes the cavity volume and uses this volume to estimate the repair cost. Thus, detection, measurement, and cost estimation are fully automated in this solution. Consequently, the road inspection process becomes more streamlined and less laborious as compared to labor-intensive traditional inspection methods. This method is more accurate for pothole detection while it allows for assessment and planning for such extensive infrastructure maintenance works. . Keywords: Pothole detection, depth estimation, Volume calculation, Repair cost estimation, Cost-effective solution