Background median on g band images is unreasonably high #55
Description
I am trying to run the pipeline on some g band images and the median background calculation is returning a value that is ~10^131 electrons. The background subtraction step therefore over-subtracts and causes everything later to fail. The raw g band images look great in ds9, I'd say the background flux is about ~1000 counts.
Just to make sure this wasn't something I broke, I double checked the pipeline on some r band images that I successfully ran last night and those still work great (background ~2000 electrons).