This seems to be the year where three-dimensional (3D) technology is starting to take hold in the security industry. Coming up next week at ASIS, two companies have let SecurityInfoWatch.com know that they will be showcasing 3D technology. The first is DVTel and the second is Feeling Software. SecurityInfoWatch.com had a chance to speak with both firms, and this is the report.
DVTel's AVT-3D technology
Pop over to booth 2851 at the ASIS tradeshow in Dallas (Oct. 12-14, 2010), and you'll find DVTel exhibiting 3D video surveillance technology. This company is probably best known for producing a line of video surveillance cameras and the associated Latitude Network Video Management System for managing a video surveillance installation.
What's new for the firm is their AVT-3D solution. AVT stands for Adaptive Visual Technology, and according to DVTel's Chief Technology Officer Ed Thompson, the AVT-3D system is built upon DVTel's SceneTracker system. The SceneTracker was a DVTel VMS solution that allowed "stitching" together of different camera views into a holistic view.
The 3D aspect of their technology is that DVTel's AVT-3D solution can tie together standard, high-definition cameras in pairs to register images into 3D. To understand this, it's best to explain what traditional 3D video is.
Three-dimensional video is typically created by what are called stereoscopic cameras (see an image of a stereoscopic camera). Basically, these are two cameras positioned close to each other (much like human eyes) and registered together so they are viewing the same scene. By having some separation between the two sensors, you're capturing two different perspectives, which when taken together can actually provide a way to determine the depth of an object in the field of view.
But since our industry uses standard surveillance cameras and not stereoscopic cameras, DVTel uses server technology to register the views of two separate HD surveillance cameras.
"SceneTracker technology allows pixel-by-pixel registration of the two images to get a very natural experience," explains Thompson. "That is the tough part of 3D if you don't have a purpose built 3D camera."
Thompson says the model is to take two standard or HD cameras co-located any distance apart (2 cm to a meter or more), and allow DVTel's technology to create an overlay instead of stitching them together. For the show, DVTel is planning to show 3D using 1080 HD cameras at 30 fps pushed to a 52-inch monitor with a 120hz update rate for viewing the scene. Glasses will be used, though Thompson notes that some monitor manufacturers are getting closer to being able to allow 3D viewing on monitors without the need for special 3D glasses.
"The closer you put the images together, the less depth perception you get," adds Thompson in explaining how the technology is set up. "If your eyes were wider apart, you could actually see things further away when you first notice they're moving, but too far apart and you would get a double image."
"I don't know that we've found all the killer apps yet [for this AVT-3D technology], but we see interesting things possible in places like airports, like concourses where there are lots of people. It really brings out the information about people standing there and separating them out of the scene."
Thompson says that video analytics could benefit tremendously from this technology. "Analytics could benefit from the 3D perspective because it can tell how big something is. In our ioimage product line [ioimage was acquired by DVTel], we have a 3D calibration which goes through to teach the camera what the 3D landscape looks like. You do two or three of these calibrations, and that helps the camera know the difference between human, small animal and a truck. In analytics, 3D could add a tremendous capability in driving false alarms down and easing deployment."