This is a very good contribution to geomorphological mapping of glacial terrain and in particular ribbed moraine which automated mapping approaches to date have overlooked.  Whilst it is clearly not perfect, which is freely acknowledged by the authors, it also clearly represents a great step in the right direction towards developing a mapping approach is in my opinion badly needed, because mapping ribbed moraines over the large areas needed to test formational theories and reconstruct ice sheet sectors is extremely time consuming work.  This approach, at least shows good results in areas where the terrain cover is mountainous and should work well in Norway and parts of Sweden.  The mapping produced obviously needs cross checking and then editing as the overall accuracy is not high enough yet, but this is likely to be  less time consuming that manual mapping of large ribbed moraine terrains that can contain 10s of thousands of individual ridges.  I have some specific comments uploaded for the authors to consider.  

I recommend this paper to be accepted in its present form.  It is clearly written and describe subject quite well.