Detection of Traffic Congestion Using Deep Learning Techniques

  • Dr. B. Sivakumar, Rituraj Kadamati, Gopinandh Chandu

Abstract

In India the traffic frequency is increasing daily and more vehicles keep increasing daily and many people waste time waiting in traffic jams. So as to take care of this issue of high traffic pressure, it is imperative to explain the traffic clog. The imagery covers a range of road configurations, times of the day, atmospheric and lighting conditions. A subset of the symbolism has been arranged by naming the pictures with an apparent traffic blockage state. This named dataset is utilized to pass on profound learning system of neurons for the objective blockage state acknowledgment issue. Here a two-stage arrange is prepared utilizing profound learning for the picture handling part and a bespoke profound subnet for the clog acknowledgment step. The subnet is prepared on delicate class marks, these considering labeller certainty. The joined classifier acquires 95% exactness on held-out test tests, with a large number of the misclassified pictures ending up being marginal cases. Different outcomes incorporate relapse style yields which rank an assortment of pictures arranged by expanding blockage.

Published
2020-05-07
How to Cite
Dr. B. Sivakumar, Rituraj Kadamati, Gopinandh Chandu. (2020). Detection of Traffic Congestion Using Deep Learning Techniques. International Journal of Advanced Science and Technology, 29(06), 3556 - 3561. Retrieved from https://sersc.org/journals/index.php/IJAST/article/view/14156