Inspiration

3D scanners are commonly used for bringing physical models into CAD environments, but they often require hours of manual work to fix them up to a usable state. We wanted to make a scanner that removes all of that hard work by directly converting images into dxf files usable in CAD.

What it does

Using images, we detect the outline and significant features of a physical object and convert it to line segments, circles, and arcs. Which can then be brought directly in cad to significantly speed up the design process.

How we built it

Using a common camera, we captured high-resolution images for computer vision processes. We then created a pipeline that detects significant features like edges, circles, or arcs using OpenCV and our own algorithms. Afterwards, we cleaned up features to create continuous boundaries and exported them to DXF.

Challenges we ran into

Getting a consistent and reliable algorithm for arc detection was surprisingly difficult, as it's a much more complex shape than line segments or circles.

Accomplishments that we're proud of

We created a full OpenCV pipeline and created a useful product for engineers and makers.

What we learned

How complex feature recognition algorithms work.

What's next for BLUEPRINTS

Spline detection and CAD program integration

Share this project:

Updates