Imaging sensors are just electronic devices that receive photons (light) and turn them in to electrons (electric signal). As such, they are affected by ambient conditions such as temperature.
For example, you will have some noise coming from the electronic agitation (false currents due to freely moving electrons, without incoming light) that grows with the sensor temperature.
You do also have non-linear responses in the sensor that cause other noise patterns. A pixel needs to receive enough light before providing a reliable measurement. Before that threshold, it will be affected by Poisson noise (a noise whose distribution depends on the value of the pixel). On the other side of the per-pixel energy, it can get saturated (too much light), and its energy will "spread" onto neighboring pixels, thus modifying their value.
Finally, since you have a digital device, then you have an analog-to-digital conversion step that is likely to introduce further quantization noise, i.e., the recorded values will not be the "real" ones but something close up to the measurement precision.
Outside your imaging device, there is light transmission in the atmosphere (turbulence is very likely to be a disturbance cause here, but it's usually not considered as noise), and image transmission over communication channels where packet losses (and thus pixel loss!) can happen.