There are several features that may be supported in network cameras to improve image quality are backlight compensation, exposure zones and wide dynamic range.
While a camera’s automatic exposure tries to get the brightness of an image to appear as the human eye would see a scene, it can be easily fooled. Strong backlight can cause objects in the foreground to be dark. Network cameras with backlight compensation strive to ignore limited areas of high illumination, just as if they were not present. It enables objects in the foreground to be seen, although the bright areas will be overexposed. Such lighting situations can also be handled by increasing the dynamic range of the camera, which is discussed below.
Besides dealing with limited areas of high illumination, a network camera’s automatic exposure must also decide what area of an image should determine the exposure value. For instance, the foreground (usually the bottom section of an image) may hold more important information than the background; for example, the sky (usually the top section of an image). The less important areas of a scene should not determine the overall exposure. In advanced Axis network cameras, the user is able to use exposure zones to select the area of a scene — center, left, right, top or bottom — that should be more correctly exposed.
Wide dynamic range
Some Axis network cameras offer wide dynamic range to handle a wide range of lighting conditions in a scene. In a scene with extremely bright and dark areas or in backlight situations where a person is in front of a bright window, a typical camera will produce an image where objects in the dark areas will hardly be visible. Wide dynamic range solves this by applying techniques, such as using different exposures for different objects in a scene, to enable objects in both bright and dark areas to be visible.
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.