The astronomy crowd fights noise by mapping the response of the entire sensor when there is no light. Then they use that map to correct the exposures made. They are relying on the sensor noise, cell by cell, being a repeatable effect. As opposed to noise that is introduced in some other part of the pipeline.