Examining the CCTV Footage from the Marriott BombingBy John Honovich, Published on Sep 21, 2008
Below is a recording of the actual CCTV footage from the Marriott bombed in Pakistan. It looks like it's from a DVR, recording about 5-8fps at CIF or 2CIF. This is solid performance for a system that was likely installed in the last few years. The question is: could we improve this with today's technologies? Watch the video and I will comment below.
The camera is well positioned, looking straight ahead and at a minimal vertical angle. The view would be moderately better if the camera was lowered but it is likely that logistical issues prevented this.
The video does clearly show the movement of people and cars. The challenge is that the field of view is over 40 feet wide. It's essentially impossible to tell the details of the terrorists, the truck or the explosion.
Megapixel Cameras: Clearly, a megapixel camera would sufficiently improve the evidentiary value of the camera. Assuming this camera was recording at 320 x 240, less than 10 pixels are available per foot. Increasing the resolution of this camera to 2 megapixel (1600 x 1200) would would provide a 500% increase in pixels per foot, providing detailed evidence of the terrorist and the truck.
Pixels per foot is a good metric to use in designing CCTV systems. For more, review our report on specifying quality video surveillance using Pixel Per Foot metrics.
A megapixel camera will likely cost $500 to $1000 more than an analog camera plus an extra $1000 for storage. Nonetheless, in high security cameras and for a large critical choke point, this should generally be a worthwhile investment.
Video Management: I do not know what video management Marriott is using but I assume it's a DVR. Hotels often use major American DVRs like Intellex. Assuming this is the case, adding megapixel cameras will be a problem. Intellex and other mainstream DVRs offer very limited upgrade support to megapixel cameras. You will have to change video management systems or record the megapixel camera on a separate PC (which is unattractive). For pointers on such migrations, read my overview on migrating from analog to IP video.
Video Analytics: Analytics probably would not have made a difference given that this was a manned post. Analytics would not have provided any earlier detection as the gate was the key barrier employed.