After investing in a state-of-the-art video surveillance system, Swinburne University of Technology wanted an efficient way to monitor their camera feeds. With hundreds of cameras, there is always more video data than the security team can easily review.
Swinburne University of Technology is a leading university in Melbourne, Australia with 23,000 students. The main campus of Swinburne University of Technology (SUT) is in Hawthorn, seven kilometres from the Melbourne CBD. SUT has five campuses across Melbourne.
The SUT security team protects students, staff, and buildings across its campuses. SUT invested in a state-of-the-art Milestone CCTV system with 650 Axis cameras. This system has excellent coverage across the campuses. This allows the security team to review situations – especially after hours.
With hundreds of cameras, there is more video-data than the security team can “live” monitor. Therefore the security team needed a new way of filtering the feeds.
The solution had to work in a very busy 24×7 campus environment. Also it had to work without complex work to update rules.
Video analytics companies over the last 5-10 years have tried to automate video monitoring. However, these systems run into difficulties with larger camera installations or new situations. Also, some of these solutions have heavy hardware needs.
iCetana’s system uses a different approach. Because iCetana uses artificial intelligence to learn what is normal, it can detect “all kinds” of abnormal events. With iCetana, you don’t need to build and maintain rules. iCetana easily scales to thousands of cameras, but uses server resources efficiently.
Swinburne installed iCetana software to monitor hotspots across all five Melbourne campuses. Since iCetana is a Milestone Solution Partner, the system integrated into Swinburne’s Milestone VMS system. iCetana’s LiveWall plugin shows any detected events on a single screen.
With LiveWall, screen operators easily monitored large numbers of cameras with a single view. The LiveWall displays only unusual, actionable events. Usually about 1% of the total video from the camera network is unusual.
This reduces the security operator’s workload. This allows them to work on other tasks and focus on unusual incidents. Many of the incidents shown were of vandalism, loitering, fights, anti-social behaviour and vehicle violations.