
IoT-Based Automatic Vehicle Counting and Traffic Forecasting
Software as a Service
NoCategory
Thesis DesignClients
Techstack
Purpose
Mapua University research applying the Internet of Things to traffic management: automatically count passing vehicles from a roadside camera and use the collected data to forecast traffic flow for better transport planning.
Description
A Raspberry Pi roadside unit that detects vehicles in the camera feed with a YOLOv5 model running on PyTorch, then counts them using centroid tracking in OpenCV. Counts are logged on a schedule and pushed to ThingSpeak and Firebase for cloud storage and live monitoring, with ngrok exposing the device for remote access and email reports keeping stakeholders updated. The accumulated time-series data feeds the traffic-forecasting analysis at the heart of the study.




