Vehicle Traffic Analysis Using Yolo
More details
Hide details
Publication date: 2019-04-24
Eurasian J Anal Chem 2018;13(Engineering and Science SP):emEJAC181256
Traffic density specifically in the crowded urban areas is at an all-time high. It requires highly accurate and fast traffic analysis systems for capturing data to produce insights and for surveillance purposes. The data of vehicle traffic collected over a time period can be used to find traffic density patterns and procure insights which can be used for improving the traffic management. Existing hardware-based techniques for traffic analysis include magnet based loop detectors embedded inside the road provide useful data, but also has a significant downside: physical damage over a period of time, which reduces their functionality and accuracy. Even most of the software based techniques perform well to an extent, however they can only detect moving vehicles. To solve this issue, in this paper proposes to use a convolutional neural networks based algorithm known as You Only Look Once (YOLO). This paper proposes to create an end-to-end traffic analysis system which can take video as the input, process the video using YOLO algorithm and produce the output report using which insightful analysis can be obtained. The data is obtained from a surveillance camera to evaluate this model.