Social distancing is essential to prevent the spread of COVID-19. The Hong Kong SAR Government has put in place measures to prohibit group gatherings of more than a certain number of persons, implemented over different periods of time. However, huge manpower or Internet Protocol (IP) cameras are required to monitor social gatherings. By applying artificial intelligence in a proprietary depth camera system, Cyberport alumnus VisionMatrix Technology Limited found a smart solution to count the number of people in an accurate and efficient way.
AI driven depth sensor ensuring non-intrusive data collection
With facial recognition capabilities, some video surveillance systems are able to accurately gather data. However, it has also created concerns about compromised privacy, or abuse of surveillance.
In view of the general public’s concern, VisionMatrix has developed a depth sensor module which combines a depth camera with a 3D AI engine. This AI-driven depth sensor can acquire multi-point distance information and obtain a precise picture with full 3D depth perception. It enables object counting and motion tracking, without collecting personal identity or capturing any details and features to preserve data security.
The depth sensor can be installed on a patrol car or a venue to monitor the number of people gathering together, or to measure the social distance of people in real time. An alarm signal will be sounded when the sensor detects anything that is beyond a certain limit, or an observation of behaviour that is forbidden by law.
Currently, a similar depth sensing module produced by other countries costs around US$3,000. However, VisionMatrix is planning to apply their depth sensing optimisation IP to a locally made product for US$500 apiece.
- Facilitate human observation with AI technologies.
- Enhanced depth camera to protect personal privacy.