Next Challenge: Human eye

A portion of the Triangulum Galaxy showing the result of stacking multiple exposures to reduce noise and increase signal.

One way to take pictures of fast moving objects is to take an extremely short exposure, freezing the movement of the object. Unfortunately dim objects in space cannot be detected easily with short exposures. Imaging dim objects requires long exposures. A pessimist might say: "Wait long enough and something bad is sure to happen!" When it comes to long exposures in astrophotography he'd be exactly right. Sometimes the "bad" is your mount having a fit. Sometimes it’s a deer running into your gear in a panic. It may even be a bird dropping a load on your telescope. If it is not one of those things, it will likely have something to do with noise. Noise is simply unwanted signal in your image. Noise in your image can come from many sources. In electronic cameras many of these sources are predictable and so they can be removed from the image mathematically through calibration (technically speaking they are not "noise"). I spend many hours doing these measurements and performing the calibration steps. There are other types of noise that are not as predictable. One is thermal noise in the camera sensor. This can be addressed by cooling the sensor. I cool mine to -20C. Other sources of noise are not so easy to deal with. These can be satellites moving across the sky, gamma rays hitting your camera, the random nature of photons travelling through space, or just a pixel in your camera imagining it saw something. The longer you do a exposure the more likely something bad will happen. 

One way to think about this is to imagine that you're travelling through a noisy city trying to listen to your favorite radio show. Cars honk, people yell, music blares from the car next to you, and you miss what was said on the radio. The longer the show, the more bits you miss. That random noise keeps you from detecting the signal you were trying to pick up. Lucky for you the show is re-broadcasted every hour. If you're stuck in your car long enough, you hear it a few times and pick up on what you missed the previous broadcasts. Chances of there being a loud noise at exactly the same time the next time you listen to the show is small.

 I deal with noise the same way in my images. I take many shots of the same portion of the sky and stack the exposures to reduce the effect of the noise. Take a look at the black and white picture of a portion of the Triangulum galaxy. One shot has a lot of noise, as I stack (integrate) more exposures the signal I want to detect increases and the noise decreases. What is powerful about this technique is that I can collect photons over many nights and integrate the exposures to generate a final image. Most of my images represent data collected on many nights, sometimes months apart.