InnoMind® Competition

Finalist Team – 人流智管

CrowdFlow Intelligence is an AI-powered crowd management system utilizing drone and deep learning technologies. We aim to address crowd congestion, sudden crowd gatherings, and delayed traffic management decisions.

CrowdFlow makes use of DJI Air 3 drones equipped with LiDAR sensors and high-resolution RGB cameras for real-time crowd density monitoring. Our Convolutional Neural Network (CNN) model is trained with 3D pedestrian and road network datasets from government open-data platforms. It is then trained and data cleansed for image analysis and pattern recognition functions.

Our key features include:

1. Real-time Monitoring:
- Track crowds and traffic density percentages
- Display data via heatmaps for intuitive visualization

2. Smart Alert System:
- Trigger warnings displayed on the dashboard when crowd density exceeds 80% capacity in a specific area

3. Generation of automated rerouting suggestions based on low-density pathways

4. Public Integration:
- Collaborate with news platforms (e.g. HK01, Yahoo) to provide real-time heatmaps and evacuation guidance to the public

Our target users are the Transport Department (for traffic flow optimization), event organizers (for on-site safety management) and the general public (for real-time crowd avoidance).

CrowdFlow's vision is to enhance efficiency in crowd control for Hong Kong's major events, ensuring safer and smoother experiences, while advancing our Smart City development objectives.

For more information, please contact dare.to.change@ydc.org.hk

Showcase