Traffic Automation Using Computer Vision
|© 2020 by IJCTT Journal|
|Year of Publication : 2020|
|Authors : Vikas P, Meera N, Lillit Francis, Vimal Mohan|
|DOI : 10.14445/22312803/IJCTT-V68I6P114|
How to Cite?
Vikas P, Meera N, Lillit Francis, Vimal Mohan, "Traffic Automation Using Computer Vision," International Journal of Computer Trends and Technology, vol. 68, no. 6, pp. 82-86, 2020. Crossref, https://doi.org/10.14445/22312803/IJCTT-V68I6P114
Traffic these days are increasing day by day and the prolonged waiting in traffic signals are getting into the nerves. The process of traffic monitoring is predominantly carried out manually in our country. This leads to many difficulties like wastage of time, higher fuel consumption, increase in pollution, higher chance of collisions between vehicles. Our project aims to completely avoid or reduce this waiting time and all the problems associated with it. Furthermore, we’re also providing an extra module which identifies the emergency vehicles stuck in traffic and allows them to move without having to wait. This will help save lives.The project mainly has two phases – Vehicle detection and Traffic scheduling. The former comprises three sub- phases that are Image acquisition, object detection and density calculation. Real-time traffic data is taken as input and is sent to the vehicle detection phase. YOLOv3 is the model used for object detection which is a single shot detector. It uses a single neural network and predicts the bounding boxes and probabilities of each region. YOLO will display the current frame and the predicted classes as well as the image with bounding boxes drawn on top of it. From this, the density of each road is obtained and is sent as the input data for the next phase. Subsequent processes are carried out in a novel Traffic Scheduling algorithm, which is similar to that of the Round Robin algorithm but with a variable quantum. Instead of turning on the green light for a fixed amount of time, the duration will be managed dynamically based on the amount of traffic on each road. This helps in solving many of the difficulties faced in conventional systems and also the data obtained from this can be used for further applications.
component, formatting, style, styling, insert
 S. M. Shinde, “Adaptive traffic light control system,” in 2017 1st International Conference on Intelligent Systems and Information Management (ICISIM). IEEE, 2017.
 L. Bhaskar, A. Sahai, D. Sinha, G. Varshney, and T. Jain, “Intelligent traffic light controller using inductive loops for vehicle detection,” in2015 1st International Conference on Next Generation Computing Technologies (NGCT). IEEE, 2015.
 L. Chen, F. Ye, Y. Ruan, H. Fan, and Q. Chen, “An algorithm for highway vehicle detection based on convolutional neural net- work,”EURASIP Journal on Image and VideoProcessing, vol. 2018, no. 1, p. 109, 2018