Implementation of Object Tracking to Detect and Count the Number of Vehicles Automatically Using the Kalman Filter and Gaussian Mixture Model Methods
Keywords:
Vehicle counting system, Object Tracking, Gaussian Mixture Model, Kalman Filter.Abstract
Traffic density can be controlled by obtaining and managing vehicle volume data on the road. In general, the process of acquiring data on the volume of vehicles passing on the highway is still carried out manually, namely by assigning several people to the field and counting each vehicle that passes, then dividing it by a certain time period. In manual calculations there are still many weaknesses such as data collection taking a long time, and the large amount of human resources required. Based on these conditions, an accurate automatic vehicle counting and detection system is needed as a traffic control monitor and traffic analyzer.Currently, a vehicle detection system has been developed using sensors, Radio Frequency Identifiers or other hardware which is integrated by software on a microcontroller and works automatically to detect speed and count the number of vehicles passing on the highway. The weakness of this detector is that it can only detect a narrow range, the system design and operations are complicated, and the operational costs are large. Based on these several things, this research was developed with a focus on designing a vehicle detection and counting system using the Kalman Filter and Gaussian Mixture Model (GMM) methods.This research intends to design and create a program that is able to identify the type of vehicle in a video input and calculate the number of vehicles detected based on their type. Users simply enter traffic recording videos into the Car Detection program. Later, the program will process the input video and produce a .txt file as output. This file contains the types of vehicles detected along with the number of vehicles detected based on their type. Using the Kalman Filter and Gaussian Mixture Model (GMM) methods, the most accurate results were obtained in the morning (lighting 10,000-25,000 lux) with an F1 Score of 0.91111, while the least accurate vehicle counts occurred in the evening (lighting 0.27 -1.0 lux) with an F1 Score of 0.16071
Downloads
References
Intel Corporation, "OpenCV," Itseez, June 2000. [Online]. Available: http://opencv.org/. [Accessed 4 December 2016].
Intel Corporation, "OpenCV," Itseez, June 2000. [Online]. Available: http://opencv.org/platforms/. [Accessed 5 December 2016].
P. Viola and M. Jones, "Rapid Object Detection using A Boosted Cascade of Simple Features," in Conference on Computer Vision and Pattern Recognition CVPR 2001, Hawaii, 2001.
P. Viola and M. J. Jones, "Robust Real-time Face Detection," International Journal of Computer Vision, vol. 57, no. 2, pp. 137-154, 2004.
R. Lienhart and J. Maydt, "An Extended Set of Haar-like Features for Rapid Object Detection," in International Conference on Image Processing 2002, New York, 2002.
Office for Mathematics, Science, and Technology Education University of Illinois, "Mean Square Error," Office for Mathematics, Science, and Technology Education (MSTE), Champaign, 2004.
C. Papageorgiou and T. Poggio, "A Trainable System for Object Detection," International Journal of Computer Vision, vol. 38, no. 1, pp. 15-33, 2000.
V. K. Narayanan, "Vision Based Robust Vehicle Detection and Tracking VIA Active Learning," University of Florida, Gainesville, 2013.
Badan Pengembangan dan Pembinaan Bahasa, "KBBI Daring," Kementerian Pendidikan dan Kebudayaan
Indonesia, 2016. [Online]. Available: https://kbbi.kemdikbud.go.id/entri/kendaraan. [Accessed 9 May 2017].
M. Oliveira and V. Santos, "Automatic Detection of Cars in Real Roads using Haar-like Features," Department of Mechanical Engineering, University of Aveiro, vol. 3810, 2008.
S. Han, Y. Han and H. Hahn, "Vehicle Detection Method using Haar-like Feature on Real Time System," World Academy of Science, Engineering and Technology, vol. 59, pp. 455-459, 2009.
S. Sivaraman and M. M. Trivedi, "A General Active- Learning Framework for on-road Vehicle Recognition and Tracking," IEEE Transactions on Intelligent Transportation Systems, vol. 11, no. 2, pp. 267-276, 2010.
T. Ball, "Coding Robin," Coding Robin, 22 July 2013. [Online]. Available: http://coding- robin.de/2013/07/22/train-your-own-opencv-haar- classifier.html. [Accessed 23 May 2017].
