New Cisco Systems IP surveillance cameras equipped for heavy-duty analytics
ATLANTA—Cisco Systems has launched new IP cameras that boast the processing power to perform more complex video and audio analytics to support applications such as license-plate recognition and detection of gunshots and unattended objects.
The video surveillance cameras—the 6500PD and 7530PD—include a digital signal processing (DSP) chip in addition to the core ARM processer found in Cisco’s standard cameras.
“(The DSP chips) have the real muscle to be able to take video, decompress it and to run image-processing algorithms that are part of video analytics to detect edges of objects and all the other sophisticated algorithms that video analytics need to detect objects, classify them, derive information from the video,” Brian Apgar, senior director of engineering with Cisco Systems, said during an interview with IWCE’s Urgent Communications.
“We’re working with other partners as well, like iOmniscient and IntuVision, to put even more complex analytics on (the cameras),” Apgar added.
An example of those advanced analytics is facial recognition in a crowd, which iOmniscient is demonstrating in Cisco’s booth at the ASIS Seminar in Atlanta this week.
Cisco also announced that the latest iteration of its video surveillance manager (VSM)—version 7.6—would be released in November.
The combination is intended to create a solution that helps users—in particular, those managing applications for remote site monitoring—manage their networks that are quickly expanding as more data sources, like cameras and sensors, are brought online.
“When we look at this from the perspective of physical security, we are seeing that the needs of our customers are rapidly evolving,” Apgar said during a recent briefing, noting that this phenomena has been observed in many industries.
“The problems that they are trying to solve are more complex, and the scale of their deployments and networks are growing very, very rapidly. Their problems have become problems of the Internet of Everything [IOE].”
With the Cisco cameras, users are alerted when an event is detected, and then only that relevant video is either transferred, whether on demand or automatically. The rest of the video—the tedium that happens between events—is saved for two or three days before falling off the camera’s memory.
Apgar said the cameras cost “a few hundred dollars” more than a more traditional camera, and some applications must be purchased separately, but Cisco expects customers to experience a cost savings as a result of the new capabilities. Specifically, bandwidth can be better managed, large amounts of dedicated server space will not be needed, and fewer personnel will be needed in control-room monitoring surveillance stations.
“These new edge-analytics capabilities give a new level of intelligence at the edge to be able to analyze data from those cameras and determine the data that’s absolutely essential to be alerted to operators back at the core of the network,” Apgar said.
Additionally, the cameras have an open application development platform and SDK for IP cameras, giving users the flexibility to add functionality as needed and turning the cameras into “intelligent sensors” capable of more than just video, according to a press release.
Apgar said two new capabilities were announced for the VSM 7.6: scalable application management at the edge and high availability in the application server. The system can support as many as 10,000 cameras in a single operations manager and more than a million cameras in a federated system, according to a product-specification sheet.