Why are Operational Requirements so important for good quality CCTV?
To design a CCTV system that is fit for purpose and will provide the necessary quality in the event of an incident it is critical that there is a discussion about the areas that are to be covered and what level of detail needs to be provided, i.e. does the end user want to identify, recognises a known person, detect an activity or just observe.
Only when this has been agreed and ideally drawn on a site plan should the discussion about the actual cameras and locations take place. As an example, an analogue camera of standard resolution can be set up with the correct lens to view only an entrance and capture identification quality images whereas conversely a HD or Megapixel camera with much greater resolution can be set up to view too large an area to be able to identify an individual.
What size area should different resolution cameras be set up to view?
Although traditional analogue cameras that have been the mainstay of the security industry for years have been blamed for the poor quality images you see in the media, even now a standard resolution camera can still be useful even for capturing identification quality images. Home Office advice states that a person should make up 100% of the image in order to provide identification quality images although this is very much also dependant on both lighting levels and the recording settings as any compression applied will deplete the image quality.
Nowadays 1080p HD cameras are widely available both using advanced analogue technology such as HD-TVi or HD-SDi using traditional coax cable or by going down the IP CCTV route using either an existing or dedicated IP network. In this instance it is more accurate to look at operational requirement in terms of pixels per face. Roughly speaking, in ideal conditions a minimum of 32 pixels horizontally per face is required for identification based on the average facial width of 16cm.
This means that for a 1080p camera which has 1,920 pixels across the image, the view at the capture line can be no more than 9.6 metres across so if for example you have a loading bay which is 12 metres across and try to get identification quality images with a 1080p camera then clearly this will not be sufficient and it will be more appropriate to look at a camera that is at least 4 Megapixel camera as this will allow identification in an area that is approximately 13 metres across.