Cong Mu*, Tamás Budavári, Department of Applied Mathematics and Statistics, Johns Hopkins University
Today there are dashboard cameras in many cars providing a video log of our drives. The recordings of work commutes and other trips could be useful for many purposes, for example, identifying unusual incidents in trips, assisting in insurance claims for accidents, etc. Our goal is to build automatic tools for analyzing these logs and annotate the video stream with relevant information. We start by studying freely available video streams of sports cars on the race tracks. In particular we first focus on estimating the lap times of these cars based purely on the video footage. Here we present our preliminary results using a number of methods and algorithms, and discuss our future work.