Treffer: Real-time traffic light control system based on image processing and object detection.

Traffic congestion causes a plethora of problems, from wasted time and fuel to air pollution. Much of the cause of congestion is due to the inadequacies in conventional traffic light systems. Therefore, this project proposes a smart traffic light control system that is able to redirect traffic based on real-time behavior using image processing and object detection techniques for vehicle detection. A Raspberry Pi and Pi Camera act as a controller to control a set of red, yellow and green LEDs simulating traffic lights. The Python programming language and OpenCV library was used for image processing. Frames from the video feed are converted into HSV, noise reduced using a median filter and morphological operations, thresholded into binary and extracted for contours. Objects are identified from the contours and counted. The system then controls the traffic light LEDs based on the number of vehicles present on each lane, making the optimal decision to reduce congestion and allowing cars on the lanes with heavier congestion to pass. The system displayed an object detection accuracy of 96.5%, 94%, 91%, 81% and 86% and in simulated normal, bright, medium, low light and fog. Furthermore, the dynamic system showed a 53% queue time reduction compared to conventional fixed-time systems. Results show that the system performs well under various levels of illumination and is feasible for practical use with existing traffic light infrastructure. [ABSTRACT FROM AUTHOR]

Copyright of AIP Conference Proceedings is the property of American Institute of Physics and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)