Muni to Install New Automatic Surveillance System


Photo by James Castañeda

Muni stations may soon have a new video surveillance system that uses a “behavioral recognition” technology to detect suspicious activities. The proposed system by BRS Labs uses software algorithm to analyze video footage and sends alerts to a human supervisor about any irregular activities, according to a CBS report. The SFMTA public bid shows that there will be up to 22 cameras installed in each of the 12 stations using this human-free system.

SFMTA spokesperson Paul Rose told us that construction will be complete at the Portals by this time next year and complete in the stations by late next year.

The Daily Mail reports that the cameras build up a memory of normal versus suspicious behavior:

The company says will put them in 12 stations with up to 22 cameras in each, bringing the total number to 288.
The cameras will be able to track up to 150 people at a time in real time and will gradually build up a ‘memory’ of suspicious behaviour to work out what is suspicious.

So how does this work? BRS Labs President John Frazzini told Fast Company:

The company’s AIsight behavior recognition product relies on 11 patents related to computer vision technology and surveillance imagery. BRS’s patents primarily deal with the intersection between computer vision and machine learning; video footage grabbed by MUNI cameras will be automatically translated into code for real-time processing. Clips of anomalous activity are dispatched to MUNI employees automatically; SMS text message alerts are also sent to staffers’ mobile phones.

The system can send out alerts such as when someone has placed an unattended package in a populated area, or when an individual is trespassing into an area that normally only vehicles travel.

Some reports have hinted at the potential “big brother” aspect of the to-be-installed security system. Should the average transit rider worry about the new cameras?

7 comments

Leave a Reply

Your email address will not be published. Required fields are marked *