I moved a chunk of the histogram/black point code out of BaseMonochromeBitmapSource and into BlackPointEstimator which makes a lot more sense. Unfortunately I had to expose three new methods and touch a bunch of files. But I did manage to reuse an array on subsequent calls, which was being allocated on every row scanned, so that might be a bit faster. It will also be easier to convert this code to JNI in the future if we want to, and/or do more advanced thresholding.
git-svn-id: http://zxing.googlecode.com/svn/trunk@945
59b500cc-1b3d-0410-9834-
0bbf25fbcc57