I have a png image that I generated from a cad drawing using edge-detection. Only problem is all the lines have the same thickness. The lines that mark the main edges I need to be thicker and more prominent. But, I also have shading-type lines in the same drawing, that I do not want to be any thicker, as then they would merge together.
I am using imagemagick, and using
-morphology dilate disk:2 worked to make all the lines thicker. However, now the shading lines are too thick and they merged into a blob.
How can I thicken only the lines that are not the shading lines? I'm thinking that the shading lines may be identifiable by their proximity to each other. Maybe I could use that as a metric to create a mask that can identify the not-shading lines for thickening. Does this sound like it could work, or is there some other suggestion? If so, how could this be done using imagemagick (or similar)?
I made a sample model to demonstrate the effects that I've described.
First, here is the line drawing without darkening (so thinner lines) (made from imagemagick's edge-detection of the exported png):
And second, here is the line drawing with darkening (so thicker lines) (the previous one but with an additional
-morphology dilate disk:2 step).
The first drawing is too faint, it hurts my eyes to look at. The second one is pretty good, almost good enough, but the shading lines (on the two crossed cylinders) are too prominent, I think. I thought I might be able to get the best of both worlds by darkening only the non-shading lines.