I've had a number of recent discussions with end users (particularly Police Deptartments and Municipalities), regarding City-wide Wireless CCTV Systems. The desire here is to protect an area in the city with a group of WIRELESS CCTV Cameras. The secondary purpose, but often becomes the biggest purpose, is to allow an officer in the police cruiser to view the video (wirelessly) as he nears the area. This is a great concept, but it's important to set expectations of what CAN be done with reliability and stability.
Consider first the amount of bandwidth a typical CCTV Camera would require to send video to a central or semi-central location. At 15 FPS and 2CIF resolution, the bitrate might be 1024 Kbps. Multiply that times 10 cameras to cover an area in the city and you might have 10Mbps pretty quickly. Now, put that on a wireless link at a maximum of 54Mbps, it seems there is no problem. Right?
Correct. No problem, that is, if that's all that will ever go across the wireless link. But what about VIEWING the video (how many connections to the cameras)? Will the cameras ever be opened to John Q. Public? If the cameras are IP, what about sending a second encoder stream to an analytics server for server based analytics, or a secondary recorder for that matter? There are many factors that can change the bandwidth requirements across the wireless network, and they should all be considered before the solution is provided.
One great way to manage the bandwidth is to separate out wireless network infrastructure for different purposes. Consider a Wireless network with directional antennae bringing the video back to a central point. Then a second Wireless MESH network with Omni-directional antennae making the video available to vehicles and other folks as permission allows. This way the stability of the wireless network is much greater, AND a network failure on the MESH side does not interfere with the recording of cameras, or vice versa. Of course this is a costly endeavor, and wireless network saturation might become an issue in some cities.
In addition, to control the wireless stream, consider products that can be implemented in a multicast function. This way, one stream is send to the network switch. The Switch then makes available many client streams to be tapped into, rather than Unicast mode in which the encoder itself must deal with the multiple clients.
In it's new line of Wireless Products (5.0), Verint offers up some new technologies and configurations, which are specifically designed for the transmission of video to a central point. Radios are preconfigured to ive preference to upstream transmission, knowing that the grand majority of data flow is upstream (the streaming video). Using SPCF, allows the Access Points to avoid any automatic reduction in bitrates due to the common "Hidden Node" issue experienced in many of today's wireless networks.
Finally, when there is a need for multiple levels of client access to the cameras (eg, police cruiser connects at 512Kbps, wireless handheld connects at 128Kbps, and a desktop computer connects at 100Mbps), the end user might consider using an application like AirVisual's Intelliviewer and Orsus' Situator to provide a centrally located video streaming server, which can receive video from encoders, IP Cameras and DVRs, and push the video back out via a separate video stream. Some of these can also provide custom mapping features, which will show zoomed in maps, sized to fit PDA or Laptop screen resolutions. This enables the officer, on foot or in a cruiser, to quickly view the video by clicking on the camera icon on a Map.
It's easy to see how technology IS capable of doing what the end user is asking, but to do it reliably, can cost a bit more than what they are expecting.
Just food for thought...
- Matt Marshall