Achieving desired video surveillance quality

Axis Camera

When designing an effective video surveillance design, a key common factor is the amount of pixels on target. Determining the budget allowed for capturing facial recognition vs. facial identification can be vital. 80 pixels per foot is continuing to emerge as the industry standard pixel count for facial identification but this number can increase drastically when certain elements enter a scene.

Perfect Pixel Count
Identification & Recognition

HD Resolutions vs Optical Zoom

720p, 1080p, 4K, 1MP, 2MP, 5MP

Do we need a higher optical zoom or a higher resolution camera? Well, that depends on what the objective is. Is situational awareness the goal, or facial identification? We have many questions to ask here and the answer in many cases will be situational and budget may have a critical impact. It is critical to remember that while an optical zoom can get us a high pixel-on-target count, without the resolution to back it we may lose critical accessory detail by having to zoom too far. Throwing either a massive lens (even when varifocal) or a 4K camera at a situation can be expensive and may still not achieve desired results. The proper lens, sensor size, frame rate and resolution are all critical aspects that can vary widely scene to scene and camera to camera, often within a manufacturer's lineup. While there are dozens of freely available tools to try and estimate the best camera for a situation, there is a fine art behind capturing forensic data. Please view the Axis published paper below regarding optic & resolution.

Optics vs. Resolution

Lighting and Placement

Safe positioning, Proper Lighting, Lens Angles

Finding proper mounting points for cameras to be out of reach is one subject, but maintaining the proper angle from the camera to a license plate for example is another. Many integrators stick cameras up on the ceiling as they provide for easy cabling routes, and often put a camera out of harms way. Unfortunately, this often times introduces an angle from the lens that will not provide the forensic detail required to identify a target. Does an 80 pixel per foot on target view do us any good if someone's bald spot is all that we can make out?

Lighting is another commonly over looked aspects of a video surveillance design during walk throughs. Although lighting may look fantastic at one time of the day, this can change drastically over a few hours or even sooner based on objects in a scene. Covering lighting issues once the sun goes down is a decent start, but what about while the sun is rising or setting for instance? This can provide a particularly tricky situation, especially if a camera can be mounted low enough to get a horizontal view of a face for example. There is no replacement for deployment experience and expertise in many of these situations. We can use technologies such as Axis' Wide Dynamic Range as seen on our "Axis Camera Features" page to help with these situations but camera placement is even more important.