Real-time Traffic Analysis Using Deep Learning Techniques And UAV Based Video

In urban environments there are daily issues of trafficcongestion which city authorities need to address. Realtimeanalysis of traffic flow information is crucial forefficiently managing urban traffic. This paper aims toconduct traffic analysis using UAV-based videos and deeplearning techniques. The road traffic video is collected byusing a position-fixed UAV. The most recent deep learningmethods are applied to identify the moving objects invideos. The relevant mobility metrics are calculated toconduct traffic analysis and measure the consequences oftraffic congestion. The proposed approach is validated withthe manual analysis results and the visualization results.The traffic analysis process is real-time in terms of the pretrainedmodel used