AI-Based Automatic Passenger Counting (APC)
The internal and externally-mounted cameras on each train capture video footage in high definition – even in low-light and high contrast environments.
When imbued with the right level of intelligence, onboard cameras such as these can support a wide range of use cases, enabling additional operational and safety functions which have historically required manual processes, or the use of disparate systems and devices, for example passenger counting.
Northern was particularly interested in assessing the feasibility of such a strategy and using AI technology to analyse anonymised onboard video for Automatic Passenger Counting (APC). Icomera developed a sophisticated AI video analytics algorithm which was fully integrated with Northern’s APC business intelligence tool.
The solution counted passengers with a verified accuracy of 98.3%.
This level of accuracy and speed of processing supports Northern in managing fleet and staff capacity planning. The ability to effectively gauge vehicle occupancy levels, when shared, can also make it easier for passengers to choose less noisy or crowded environments to travel in, ensuring that they are comfortable throughout their journey.