Apparently, depends heavily on sensor design and methods-employed, and they like to keep them in a specific range.
Avoids ‘warm-up” or ‘cool-down’ time and they already know how the sensor performs under pre-calibrated temps.

I’ve always heard CCDs tend to have more noise in the image when cold. Why this is I have no idea, but it may be relevant here.

Also, HiRise is a reflecting telescope, not just a camera. I don’t have time at the moment to try and find it, but I’m about 95% sure I recall coming across a NASA document or article of some sort involving troubleshooting periodically blurry imagery from the camera in its early days. The culprit turned out to be the secondary mirror deforming slightly due to thermal contraction and making the instrument come of proper focus. Thus, they try to keep the system warm to prevent this. Again, this is just my recollection, but it makes sense on its face that thermal expansion / contraction would mess with high precision optics.

Andi: I can’t be more specific. This is what one of the members of the camera team told me when I first asked about “terrain sample” images. Nor does it surprise me, considering the large temperature ranges seen in space.