Authors :
Amit Kumar, Arpana Alka, Jyoti Prakash Sahoo, Nimai Chand Das Adhikari
Volume/Issue :
Volume 5 - 2020, Issue 6 - June
Google Scholar :
http://bitly.ws/9nMw
Scribd :
https://bit.ly/3djDko8
Abstract :
One of the most important problems every
country is facing to solve is the immense addition of the
traffic to increase burden if smooth management. This
research revolves around solving and maintaining the
traffic solution like congestion, idle road, emergency etc.
by using the help of traffic surveillance camera. In this
research, we are trying to make an efficient object
tracking solution called as Class Wise Object Tracking
Algorithm (CWOT). The conventional SORT (Simple
Online Realtime Tracking) does not take class IDs of
detected objects which the proposed methodology takes
into account. This methodology describes the detection,
accelerating speed of the detection, propagating object
states into the future frames and then associating
current detections with the existing objects using SORT
(Simple Online Realtime Tracking). Then generated IDs
via SORT are tagged to object detection class IDs using
this algorithm. YOLOv3 is used for the different object
detection in real-time.
Keywords :
YOLOv3, SORT, MOT, Object Detection, Multi-Object Tracking, Smart Traffic Management Systems
One of the most important problems every
country is facing to solve is the immense addition of the
traffic to increase burden if smooth management. This
research revolves around solving and maintaining the
traffic solution like congestion, idle road, emergency etc.
by using the help of traffic surveillance camera. In this
research, we are trying to make an efficient object
tracking solution called as Class Wise Object Tracking
Algorithm (CWOT). The conventional SORT (Simple
Online Realtime Tracking) does not take class IDs of
detected objects which the proposed methodology takes
into account. This methodology describes the detection,
accelerating speed of the detection, propagating object
states into the future frames and then associating
current detections with the existing objects using SORT
(Simple Online Realtime Tracking). Then generated IDs
via SORT are tagged to object detection class IDs using
this algorithm. YOLOv3 is used for the different object
detection in real-time.
Keywords :
YOLOv3, SORT, MOT, Object Detection, Multi-Object Tracking, Smart Traffic Management Systems
