Automatic Passenger Counting System (APCS) Using AI and Image Processing

Accurate passenger data is essential for effective transport planning, but it must also be practical to deploy at scale. Chartr's Automatic Passenger Counting System (APCS) is built around this principle.

Cameras installed inside buses capture images at fixed time intervals. These images are securely transmitted to a backend server, where Chartr's server-side AI pipeline processes them. Vision models and LLM-based analysis extract passenger counts from each image and map them to trips, routes, and time windows.

By keeping hardware lightweight and intelligence centralized, the system scales easily across large fleets. Model improvements are deployed at the server level without requiring changes on buses.

The resulting data helps transport authorities identify crowded route segments, analyze demand patterns, detect anomalies that may indicate fare leakage, and optimize schedules using real usage data.