- Smart Traffic Light Optimization

Overview

For my master’s capstone, I developed a system using YOLOv8 to detect the number of vehicles stopped at traffic lights and calculate the ideal green light duration. The system is designed to improve traffic flow and reduce waiting times, combining real-time computer vision with algorithmic optimization.

To pitch the project, I created a short artistic video that visualizes the concept. In one part, two consecutive shots show a driver tapping his steering wheel in frustration at a long red light, immediately followed by a close-up of a programmer’s fingers typing at a keyboard. The contrasting scenes communicate cause and effect and reflect the human versus technological perspective, while the sequence highlights tension and resolution in a visually engaging way.

Another sequence features a distant shot of a traffic light, immediately followed by a laptop screen running the code. The reflection of the red circular light on the screen subtly connects the real-world traffic signal with the program controlling it. This choice bridges the physical and digital worlds, giving the viewer a sense of the system’s direct impact.

This following scene is a 3D waveform animation created in Blender, representing the traffic light scanning the cars waiting at the intersection. The visualization abstracts the data flow into a dynamic, visual form, showing how information is processed and decisions are made in real time.

Tools Used