Realtime Cars Counting System for Smart Traffic Light System

  • Kurniawan A
  • Zaini A.

Abstract

Many cities in the world nowadays are suffering from traffic jam and traffic density due the
inevitable growth of vehicles on the street. This phenomenon is not only happening in a
modern countries, but also in many developing countries, like Indonesia. Some designs of
smart traffic lights systems have been introduced by many researchers and also
implemented previously. However, the real condition of each city is unique and need a
specific solution for their traffic jam condition. This paper presents a design system and
implementation of automatic car counting, as part of smart traffic lights system in an
intersection. The system consists of IP camera and computer in a road. Video and Image
processing technique were developed to capture and identfying the vehicles and counting
them and then stored the information in a server. Two methods of image processing
technique in counting the vehicles were implemented to get the best result. They were
square pixels ROI (PR) and Countur (CO). The experiment was done during day and night
using 3 different video resolution and 4 different frame per second (fps) technique. The
result showed that CO method showed a more robust result compared to PR, especially
during the night. Highest resolution still showed the best result for counting the vehicles.

How to Cite
A, Kurniawan; A., Zaini. Realtime Cars Counting System for Smart Traffic Light System. Proceeding ICIRAD, [S.l.], v. 1, n. 1, nov. 2015. Available at: <https://eproceeding.undiksha.ac.id/index.php/icirad/article/view/133>. Date accessed: 21 oct. 2020.