Check SNR on first section and issue warning if it seems low

[raacampbell]

It would be useful if the user had an indication of whether or not the SNR is low. However, I have noticed that my current SNR calc is odd. For AF_FL05 the SNR is reported as about 1.5 but the sample looks good. So I need a better way of judging.

I am suspicious of the SNR. We need to really explore where it comes from.

[raacampbell]

Of possible relevance to the weird SNR results is that I see we nowhere use the foreground stats produced by boundingBoxesFromLastSection: meanForeground, medianForeground, stdForeground