Testing Videofied's Video Security SystemBy IPVM Team, Published on Jan 24, 2010
Videofied has gained significant attention for tightly integrating video monitoring/recording into an intrusion detection system. The most commonly cited differentiator is the lack of any wires for any of the equipment. All of the equipment (the cameras/detectors as well as the controller) run on batteries and communicate via wireless networks.
In this report, we share test results of the V6000GPRS controller used with an Indoor and an Outdoor MotionViewer. We monitored alarms using Videofied's Frontel software that is integrated with a significant number of CMS systems.
Our key findings include:
- On-site installation and setup was simple
- Video quality was sufficient for detecting intruders but generally not sufficient for identifying small details of intruders
- Video monitoring was sufficient for dispatching responders but not fast enough for real time tracking or instant alerts
- With line of sight conditions, we achieved only a fraction of the maximum range specified (200 meters) and significantly less through walls
Additionally, we think this is an interesting substitute for video analytics in some situations with clear trade-offs (positive and negative) between the two.
We plan to test more systems for remote monitoring as part of our video surveillance test program [link no longer available].
The Videofied solution consists of 2 fundamental components:
- A panel that communicates with sensors on-site using 900 MHz and a central station using a GPRS modem. The panel is battery powered and specified to operate up to 4 years on 1 set of batteries. The MSRP of a commercial panel is approximately $750 USD.
- Video sensors, called MotionViewers that run on batteries (specified up to 2 years) and communicate with a panel on-site. Using a PIR detector, alerts trigger an on-board camera. The outdoor model specifies a black white camera with 320 x 340 resolution and an 85 degree FOV. Motion detection is specified up to a 12m/40ft wide FOV. Outdoor models have an MSRP of approximately $600 USD. MotionViewers can be placed up to 2000 ft from the panel. One panel/modem can support up to 24 MotionViewers.
Videofied reports integration with Dice, MAS, Bold, Immix, SAS, Sureview monitoring software. Detection is optimized primarily for specific choke points or assets rather than long fence lines (given the width of area detected).
The system does not allow on-demand 'live' viewing of video nor does it allow end users to access video. Video is generated as part of an alarm.
We found the following metrics for Videofied's performance:
- Maximum distance detected from MotionViewer detector/camera: 12 meters / 40 feet. This provides a maxium coverage width of 16.5 meters / 54 feet. At 30 feet from MotionViewer, detection was extremely accurate - night and day. At 50 feet from MotionViewer, detection was quite poor - especially at night.
- Delay between event detected and received by Frontel monitoring software: 45 - 49 seconds. This is a function of the GPRS connection and the transfer of video from the MotionViewer to the controller. Because of this delay, it is not possible to track intruders in 'real time.'
- Maximum distance between MotionViewer detectors and controller with clear line of sight: approximately 150 meters / 500 feet. Going through multiple concrete walls (in a multi-story building), distance was very limited - perhaps 100 feet point to point (through 3 concrete walls).
- Accuracy detection: 0 false alerts in 12 hours outdoor testing with wind/vegetation and no other false alerts during multiple weeks of general testing. We thought the system was quite accurate with no need for configuration or optimization.
These metrics are important in assessing the use and deploymet of Videofied's system.
This screencast examines each of the Videofied components and their physical elements.
The configuration screencast below examines in detail the process of system setup by an installation technician on-site.
Monitoring / Usability
Video Quality Analysis
Download the Videofied video sample zip file containing 9 videos of various day and night test scenes. In the screencast below, we examine the video quality with a focus on determining how much detection vs identification can be accomplished with this video.
Contrast with Video Analytics
We think Videofied is an interesting option for applications that need alerting and real time response of relatively small areas. Unlike video analytic systems, that generally require on-site optimization and custom integration with central monitoring stations, Videofied can be dropped in by security technicians with immediate integration into 'traditional' alarm monitoring workflows. Also, the solution is likely to be cheaper as the devices are fairly inexpensive and the installation costs for Videofied are lower (no cabling, etc.).
The most critical constraints are (1) the narrower range, (2) lower video quality and (3) delay in transmission. The detector only covers about 10 meters or 40 feet from their installation point. While this is sufficient for entrances and driveways, this would be unrealistic for parking lots or longer perimeters. Moreover, a 'true' video surveillance/analytics system would likely provide better low light performance and higher resolution. Finally, integrated with a broadband connection would trigger nearly instant alerts and allow for real time tracking of suspects. However, this would cost significantly more to accomplish. As such, these are not direct substitutes but alternatives for different applications.
Videofied provides simple deployment of integrated video verification and intrusion detection. It should not be used as a video surveillance system but it did provide sufficient video for verifying detections.
Because of its integration with central station software, it provides more of a turn-key solution than typical video surveillance monitoring options. Also, because of its use of more mature detection technologies, in small to moderate areas, this could be an attractive option to deploying video analytics.