Abstract- Vehicle growth in Indonesia is not supported by the number of the road. This fact has caused traffic congestion easily occurred, especially in big cities. Intelligent Transportation System (ITS) is one of the solution for this problem. In ITS, vehicle tracking and counting is a challenging issue for traffic surveillance. In this paper, Bhattacharyya distance-based tracking and vehicle counting is proposed. We propose a vehicle tracking method and vehicle counting system. We use Adaptive Background learning with hoked filling Algorithm to separate background and foreground in a frame. We detect objects in foreground region using contour of each object. Each detected object tracked using Bhattacharyya distance. Each tracked object is counted when the tracked object passed a predefined counting line. The proposed algorithm is evaluated by comparing tracking estimates with ground truth. We use precision and recall as metrics counting results also evaluated by comparing to ground truth. Our tracking result have 1.3% and 7.4% error rate in our counting result compared to ground ttruth.
|
|