Intelligent Traffic Signal Optimization and Vehicle Counting Using Canny Edge Detection and YOLOv8 Model

Transportation congestion in urban areas persists as a leading problem, increasing the travel time, fuel consumption and air quality. An intelligent traffic signal control system that optimizes traffic flow based on real time vehicle density estimation and counting is proposed in this paper. For vehicle detection and counting, the system relies on Canny edge detection for vehicle density estimation, and YOLOv8, a state of the art object detection model. The proposed system attempts to alleviate congestion and achieve a higher traffic efficiency by dynamically adjusting traffic signal timings with real time traffic conditions. Realspace traffic video datasets are used to train and evaluate robust performance in diverse environments. Experimental results are shown which confirm the system's ability to accommodate differential traffic densities and hence, better traffic flow and wait times than traditional fixed time control systems.