Vehicle Detection For Traffic Management System By Using Gaussian Mixture Model

Abstract

Nowadays, vehicle density has increased significantly around the world. Poor traffic management and excessive demand had caused traffic congestion in urban countries. Therefore, Intelligent Transportation Systems (ITS) play an essential role in the traffic management system, and vehicle detection system is significant in this system. The aim of this research is to investigate the efficiency of the vehicle detection system by using Gaussian Mixture Model (GMM). In this research, our objectives are to propose a vehicle detection algorithm based on Gaussian Mixture Model using MATLAB programming and examine the performance of the vehicle detection model by using real-life datasets. Since Intelligent Transportation Systems (ITS) is still not widely pervasive in the traffic network, it is vital to study the current vehicle detection system. GMM provides background subtraction for object detection in machine learning such as MATLAB and Python. In this study, the video samples are taken from SENA Traffic Systems Sdn. Bhd. and had been tested by using the GMM package provided in the MATLAB software. GMM package algorithm in MATLAB provided background subtraction for the video frame. The parameters of the vehicle detection algorithm in the computer program have to adjust for each video sample in order to obtain the better result. The results had been tested with 4 types of road video condition which are Normal Day, Normal Night, Raining Day and Raining Night in the same location, to determine the efficiency of the vehicle detection system. In conclusion, the system had come out with different results according to the different condition of road.