Privacy-preserving area monitoring with 3D cameras
GNS Engineering Team 1 min read
Cameras in public spaces raise legitimate privacy concerns. GNS Technology developed a 3D camera monitoring solution that solves this problem by design: it captures depth maps using reflected light instead of recording optical images.
The technology
Using reflected light, the camera extracts a depth map of objects for analysis. Because no optical image is captured or recorded — only the depth map — the solution is privacy-clean while still enabling powerful analytics.
Key advantages
- Designed for outdoor environments
- No optical image captured or recorded; only depth maps, giving privacy-clean solutions
- Lower cost than radar-based solutions
- Distinguishes object types: vehicles, people counting, types of obstruction
- Covers over 60 × 60 square metres with less than 10% dimension error
- Measures the speed and size of small packages and moving objects such as vehicles and trolleys
- In measurement mode, accurately measures package dimensions from 10 cm to 900 cm and extracts a 3D model within one second
Applications
- City management
- Housing and population monitoring
- Transport and road traffic status monitoring
- Passenger counting in local facilities
Learn more
This solution is registered on the Smart LAB platform as reference S-0233: Area monitoring solution using 3D camera with privacy-preserving.
Related reading
- Railway AI forecasting registered on the EMSD I&T platform.
- How we select providers: choosing a systems integrator in Hong Kong and choosing an AI solution provider.
- Our AI & IoT solutions, IoT hardware and project references.
FAQ
How does this 3D solution protect privacy?
The camera uses reflected light to extract depth maps and never captures or records optical images. Analytics run on the depth data only, so the solution is privacy-clean by design while remaining capable of people counting, vehicle classification and measurement.
What can the system measure and monitor?
It distinguishes object types, counts people and vehicles, monitors obstructions, and measures speed and dimensions. It covers over 60 × 60 square metres with less than 10% dimension error, and can extract a 3D model of packages from 10 cm to 900 cm within one second.
Where is this solution typically used?
Typical applications include city management, housing and population monitoring, transport and road traffic status, passenger counting in local facilities, and other public or industrial sites where privacy compliance matters.
How can this project be verified publicly?
The solution is registered on the government Smart LAB platform as reference S-0233, where the solution description is published for public verification.
How does it relate to GNS's other AI work?
It is part of the same engineering team that delivers railway AI forecasting, real-time analytics over 100 million train logs a day and predictive maintenance — all designed with privacy, security and production reliability in mind.
Is it more cost-effective than alternatives?
Because it does not require radar hardware, the 3D depth-map approach is lower cost than radar-based solutions while delivering privacy-preserving analytics outdoors.
