Industry
Transport
Technology stack
Computer Vision
Edge Computing
Machine Learning
Industry
Transport
Technology stack
Computer Vision
Edge Computing
Machine Learning
Industry
Transport
Technology stack
Computer Vision
Edge Computing
Machine Learning
The challenge
A public transport company had limited insights into the number of people in or around Public Transport facilities such as stops and stations due to the constant flow of patrons.
This constrained their ability to plan, respond, and generally maintain the efficiency and safety of the network in the event of routine works and during incidents.
The solution
Blackbook was engaged to investigate technical solutions, including custom hardware and software, to utilise machine vision and machine learning methods to count people and generate timely insights to positively influence safety, efficiency, and incident response.
Key components of solution:
- Utilised machine vision and machine learning methods to count people and generate timely insights
- Dedicated Edge Computing devices were installed to count people and return point-in-time counts
- Batched CCTV allowed for 15 different computer vision models to be trialed, and performances compared
The outcomes
Developed devices are capable of counting people traffic at bus stations on the edge and feeding this data back for storage and display. Infrared technology was able to count people in dark/shadowed areas, and overlapping object detection and crowd density mapping methods were able to complement each other. Having this information uploaded into dashboards provided users with current data including peak demand times at stations in order to proactively plan and rapidly respond to any situation.
