Detection of Traffic Light using Machine Vision for Autonomous Vehicles Application
Mohamed Yusof Bin Radzak1, Nur Hanim Suraya Bt. Sabarudin2, MohdFauzi Bin Alias3
1Mohamed Yusof Bin Radzak, Electrical, Electronics, Universiti Kuala Lumpur, Malaysia.
2Nur Hanim Suraya Bt. Sabarudin, Pursuing B. Eng Technology, Mechatronics, Universiti Kuala Lumpur, Malaysia.
3Mohd Fauzi Bin Alias, Lecturer, Electrical, Electronics, Universiti Kuala Lumpur, Malaysia.
Manuscript received on November 20, 2019. | Revised Manuscript received on December 15, 2019. | Manuscript published on December 30, 2019. | PP: 1033-1037 | Volume-9 Issue-2, December, 2019. | Retrieval Number: B3302129219/2020©BEIESP | DOI: 10.35940/ijeat.B3302.129219
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Traffic light detection is crucial to decrease the traffic light accidents at intersections and to realize autonomous driving. There are so many existing methods to detect traffic light. However, these approaches have several limitations, such as not function well in complex driving environments. Hence, to overcome such constraints, the traffic light detection for the autonomous vehicle using image processing technique is proposed. The experiments are carried out using 114 scene images that consist of 209 traffic lights with different angles, weather conditions, and distance. An image processing technique, Hough Circle Transform is used in this traffic light detection system with the help of Gaussian blurring and Sobel filter. So, the overall accuracy rate for the proposed algorithm is 75.59%. This system is possible to be used in urban areas or complex environments, whether it is at night or day, and it able to detect the traffic light regardless of the colour changes.
Keywords: Autonomous vehicle, Hough circle transform, Image processing technique, Traffic light detection.