Noise is considered to be the random variation in brightness or color in digital images. A pixel is a point of color and a set of pixels make up an image. Noise arises when the color and light of any of the pixels appear altered, thus spoiling the image. In analog images, graininess was accepted and was even part of aesthetics, but in digital images that noise is annoying. When we enlarge an image to see it in detail, we can see the altered pixels, especially in the darkest fringes.
Therefore digital noise we can say that it is the legacy of analog grain. This largely depends on the amount of light in the scene and the ISO we are using.
Types of noise in an image
Within the noise of a digital image we can find different types of noise, below we will detail those that normally appear.
- Color noise. Within this type, what varies is the color and is mainly related to the temperature of the camera sensor.
- Luminance noise. It is the most common, in it appear pixels with light or brightness data different from what they should have.
- Impulse noise. It is the type of noise that produces white dots on black backgrounds and black dots on white backgrounds, they appear when the sensor is worn or damaged.
- Gaussian noise. It’s not really a type of noise but rather a type of distribution that it has. Thanks to editing programs that recognize Gaussian noise, it is easier to suppress.
- Shot noise. It occurs when less light is captured in a specific area with respect to the brightness of the rest.
- Dark current noise. It is the noise that is shown due to the heating of the sensor either by hours of use of the camera or by the temperature.