[Plugin] Pic2Shape - v1.1a - 20 Mar 25 (Image Contouring)

rv1974

There's a serious problem (based on what I saw in YT video) with jaggedness. the plugin needs some contour refining\smoothing\straightening engine.

fredo6

You can do smoothing of curves afterwards, with dedicated tools. It's difficult to do it appropriately when extracting contours from the image, because there are too many situations.

And of course, without the plugin, you can always draw over the image with arcs, splines and lines....

TIG

Feel free to cannibalize parts of my ImageTrimmer, which smooths jaggedness...

You pass the edges' collected vertex points, and set an epsilon factor... thus.

      ### epsilon is a float perhaps < 0.2 
      ### perhaps it's set by the user ? experiment for best default...
      ### the vertices are collected from the current outline[s]
      points = vertices.collect{|v| v.position }
      simplified_curve = douglas_peucker(points, epsilon) ### might return []
      simplified_curve << simplified_curve[0] if simplified_curve[0]  ### so it 'loops'
      edges = some_entities.add_curve(simplified_curve) if simplified_curve[0]
      ### sort out edges' faces, hide edges etc as desired

the 'douglas_peucker()' method is this...

  ### http://en.wikipedia.org/wiki/Ramer-Douglas-Peucker_algorithm
  ### http://en.wiki.mcneel.com/default.aspx/McNeel/
                                        PolylineSimplification.html
  ###
  def douglas_peucker(points, epsilon)
    ###
    return points if points.length < 3
    ### Find the point with the maximum distance
    dmax = 0
    index = 0
    line = [points.first,  points.last]
    1.upto(points.length - 2) { |i|
       d = points[i].distance_to_line(line)
       if d > dmax
         index = i
         dmax = d
       end
    }
    ### If max distance is greater than epsilon, recursively simplify
    result = []
    if dmax >= epsilon
      ### Recursive call
      recResults1 = ImageTrimmer.douglas_peucker(points[0..index], epsilon)
      recResults2 = ImageTrimmer.douglas_peucker(points[index...points.length], epsilon)
      ### Build the result list
      result = recResults1[0...-1] + recResults2
    else
      result = [points.first, points.last]
    end#if
    ###
    return result
    ###
  end #def

fredo6

Thanks TIG.
Pic2Shape already includes a simplify method, more sophisticated that the Douglas Peucker algorithm.
The real issue is that the smooting is a matter of mixing local and global scale, but also boundaries of pixels, which itself is not always clean.

rv1974

it'd be interesting to process this fence in free inkscape for comparison

fredo6

Please do it and report.

DXF conversion is likely to give better results.

kimga

I was about to try this extension out - using SketchUp 2025 Pro, but the extension doesn’t seem to process the image. Image size 155x50 mm / 300 pixel/inch.

After 10 minutes still at Initialization…

Is there any guideline on max/ recommended picture size?

I tried it with a smaller resolution image, but I still can not get through Initialization....

Thank you

kimga

update: I succeeded with the plugin, when I opened the image in a new file.

The original image was on a bent plane. I suppose that was the problem…

pic2shape 2.jpg

rv1974

@kimga Are you going to to have trees with this amount of edges per leaf ) ?

nasmrtcz

@kimga @TIG Hi. I also have this issue. It works really great in the picture, but after applying it on a textured face, it freezes like the leaf.

fredo6

Do you have the model for me to explore the issue.

schramdesigns

I installed Fredo6 Pic2Shape but can not find it anywhere to use it. What gives?

Rich O Brien

@schramdesigns Have you met the requirements in the top post?

schramdesigns

I installed this and cant find it to use it.

Rich O Brien

That suggests you either haven't installed Libfredo or you have not restarted SketchUp. But since you are omitting lots of info to help us understand fully its hard to say.