Any image or video will contain a certain amount of noise, i.e. pixel values that are not correct representations of the scene and that negatively impact the viewing experience, aesthetically or practically. The noise is picked up from a variety of sources including the image sensor and the other electronic components of the camera, but also from the light itself.
Different factors work in favor or against reduced noise levels. For example, high resolution cameras typically have smaller pixels in their image sensors than standard resolution cameras. Smaller pixels are more sensitive to noise, in turn making megapixel and HDTV cameras more sensitive.
To reduce noise is thus a key task in a video surveillance camera. Modern network cameras that offer high processing capacity in their chips are well equipped to analyze and reduce noise levels, even in high resolution and full frame rate conditions. One technique for minimizing noise is built on spatial processing, where a single image frame is analyzed to find pixels that are very different in color or intensity from their surrounding pixels. Another technique is temporal processing, where consecutive image frames are compared to find artifacts in the images that are not static over time and can be regarded as potential noise.