| I&T Solution |
Video Analytic for People Crowdedness
(REF: S-1515) |
| Trial Project |
|
| Solution Feature |
- Analysing crowdedness of a large scale event from video data collected from All-in-one AI CCTV cameras that are catered for short term venues – battery powered that is simple to commission/decommission.
- An AI model for people counting will be used to analyse video stream, the AI model can be fused into the stream and return video stream with people count and people detected
- Occlusion, different weather, and background can be handled by the CCTV camera and AI model. Sensor is not required although it can be used as a double confirmation of the people count from the AI
- A webUI which has a dashboard to view the crowdedness with different colour indicators, and the people count of the overhead sensor.
- The WebUI also includes a management page where coordinator can manually adjust the crowdedness status, an account management module, and a report page showing the analytics.
|
| Trial Application and Expected Outcome |
- The trial will be conducted by accessing the videos that the CCTV collected from the venue and transferred back to the cloud environment in Hong Kong.
- When gained access of the video stream, AI will analyse the count of people and to determine the crowdedness level based on the customised threshold. The crowdedness level and the people count result will then be sent to the UI via API.
- The data will be updated by a 5-10 minutes interval. The webUI and parties who have the access to the API would be able to view the crowdedness level and also the people count updates and to determine if there are any further actions.
- There will be dashboard showing the crowdedness level of different regions in the map. The venue coordinator can manually override the result of the status. There will also be a reporting function showing status change overtime for analytic purposes.
- With successful outcome of this trial, this solution can be expanded to different large events held in Hong Kong. The accurate status for different areas of the event can be shown in the government page for citizen's reference and leading to better crowd management.
|
| Additional Solution Information |
SaferAI-5GBCBE-Flyer.pdf
|
| Info on I&T Solution Provider |
| Solution Provider | : | Karta-X Technologies Limited | | Address | : | Unit 316, 3/F, Building 19W, 19 Science Park West Avenue, Hong Kong Science Park | | Contact Person | : | Paau Ying Ren |
| Position | : | CEO | | Tel | : | 36190620 | | Email | : |
projects@karta-x.com |
|
