Introduction

Introduction
The consumer camera and mobile phone market focuses on image resolution as the main indicator of quality. As a result, the megapixel benchmark has carried across into the surveillance industry, and this means that network cameras are being selected based on only the narrowest criteria. This may render the camera incompatible with its functions, and may not provide the image quality required by the user.

More megapixels don’t necessarily mean clearer and more usable images. This is especially true in ultra
low or high-level light, conditions for which cameras must be optimised. Megapixel cameras give varied
results depending on what they are used for, and so, contrary to popular belief, high-quality images cannot be guaranteed. This means that security cameras are likely to produce lower-resolution images than expected, making them less usable for activities such as face or registration plate recognition.



Fig 1. Goal definition: is it enough to detect people or do you need to identify faces?

All this talk about image quality – isn’t usability the real issue?
By far the most important consideration when purchasing a new surveillance camera is that the images it produces must be clear and usable – and don’t merely satisfy an arbitrary technical requirement. Users of surveillance cameras need the best image quality out there for their specific usages, and have been led to believe that the best is represented merely by pixel number. This just isn’t the case.