In a previous post, we looked at some footage published where the police were looking for help to identify someone in a car park.
The images were of no use for the intent of the coverage as they are all too wide.
After modelling it in a 3D CCTV CAD, we get a coverage map as follows, with each camera having a red area indicating that ID quality will be achieved in that area.
The 4 resulting megapixel camera images show that we now have acceptable video surveillance that can actually be used for identification. We also have significant overlap of coverage around the entry door, giving different angles, making person ID even more effective.
To demonstrate the steps involved in making this 3D CAD model, we have recorded a video, which is at the bottom of this page.
Tim, thanks for the share…i think most of us makes incorrect analysis while putting CCTV which renders them literally worthless in times of real need…a bit of planning and strategic advice does goes a long way to make these devices really effective…
I spend a significant amount of time helping people fix their CCTV systems after they paid someone good money to install in in faith that they were getting what the needed to protect their business.
The most common issue is trying to get a camera to cover too much.
EG – a small shop wants to cover its shops shelves AND identify the person coming in the door. Reality is in most installations this is 2 or more cameras. One for ID and the other for the overview.
Another misconception is that a Mega-pixel CCTV camera will fix everything. The reality is to get double the resolution, you need 4X the number of pixels.
EG – 4CIF has 704×576 pixels (or 4SIF in some parts of the world), which is 0.4 Mega-Pixels.
Double the resolution is not 0.8 Mega-Pixels but rather 2×704 by 2×576 = 1408×1152, being 1.6 Mega-Pixels.
So to get one camera to do the job of 2, we may need to jump up 4 or 8 times the number of Megapixels.