An Improved Smart Traffic Signal using Computer Vision and Artificial Intelligence

Abstract

The growth in population all over the world and in particular in India causes an increase in the number of vehicles which, create complications regarding traffic jam and traffic safety. The primary solution to recover the jam condition is the expansion of capacities of roads by building new streets. However, this requires extra efforts and more time that is a costly and ineffective solution. Therefore, there is a need for alternative solution methodologies that are being implemented. Intelligent traffic monitoring is a branch of intelligent transportation systems that focuses on improving traffic signal conditions. The key goal of such an intelligent monitoring system is to improve the traffic system in a way that reduces delays. Many cities facing these delays because of the inefficient configuration of traffic light systems which are mostly fixed-cycle protocol based. Therefore, there is a profound need to improve and automate these traffic light systems. The establishment of a mixed technique of artificial intelligence (AI) and computer vision (CV) can be desirable to develop an authenticated and scalable traffic system which can aid to solve such problems. Proposed work supports the use of computer vision technology to build a resource-efficient, synchronous and automated traffic analysis. Video samples were collected from multiple areas to use in the system. The system applied and the vehicle was counted and classified into different classes. Manually and automatically annotated patterns were used for the classification. The multi-reference-line mechanism employed to find the speed of the vehicle and analyze traffic. The system makes its decision based on a number of vehicles, backwards-forward synchronous data and emergency conditions.

Performance Optimization of Adaptive Traffic Lights Using Machine Vision

“…Ряд работ посвящен применению компьютерного зрения в целях повышения пропускной способности перекрестков [9][10][11][12].…”

“…В работе авторы Mishra S., Birchha V. представили систему, использующую искусственный интеллект и компьютерное зрение для сбора и анализа данных о дорожном трафике [9]. В исследовании авторов Narnolia V., Jana U., Chattopadhyay S. и Roy S. представлен алгоритм интеллектуального регулирования трафика, который будет учитывать плотность дорожного трафика в реальном времени с использованием обработки изображений по каждой полосе движения, чтобы сократить время ожидания и улучшить общую пропускную способность [10].…”