In my previous article I described using 123D Catch to create a 3D model from ordinary pictures. Here’s how you can manually fix stitch pictures together into the 3D model when 123D Catch fails to do so automatically.

My project (which is taking far too long) is to create a 3D printed replacement part for a patio chair. The part is a plastic foot which has gone missing. The challenge is the very complex shape of the part. What you see in the photo collage above is using 123D Catch to assemble a 3D model from 40 photographs taken from all sides of the object.

In my previous efforts, I was not getting enough resolution of the side extrusions, so I flipped the part vertically. This allows more light to penetrate the grooves between the side extrusions. I was careful to obtain good ambient lighting (cloudy but bright day) and also left the red markers in place. Also, I played around with the camera to get better focus.

This fresh set of pictures provided much better detail. But I once again had the problem that using 123D Catch only produced one half of the model and ignored one hemisphere of pictures. This was described in my previous article.

The top two pictures above show using 123D Catch in manual “stitching” mode. This is where you can try to stitch together pictures into your 3D model where the software fails to do so. The idea is to manually place four exactly matching points on a “missing picture” as well as two “included pictures”. In the top row above, the picture on the left side is one that 123D catch did not include. The two pictures on the right side are already included in the 3D model.

This is where my little red duct tape markers came in handy, as they provided sharp corners on which to place matching points on the three pictures. I selected as my “missing picture” one that was on the edge between the two hemispheres. Once I had placed the four sets of markers, I submitted the project back to 123D Catch for attempted repair. The software tells you when you have good matches as the points turn “green”.

Using 123D Catch – First Decent Model

When “the cloud” returned my repaired model, I was thrilled. It actually worked. Manually stitching in the corner piece between the two sides or hemispheres of the part provided 123D Catch with enough information to also include all the missing pictures. The result was a 3D model with reasonably good detail created using 123D Catch. See the bottom left picture above.

I figured this model was simply as good as it was going to get using 123D Catch. So I saved the model as what is called an Object file, and opened it in Meshmixer. And there, in the bottom right picture, is the 3D model ready to print. Meshmixer automatically performed some repairs to close holes in the model. If I can figure out how use Meshmixer properly, I will try to do some more repairs and smoothing of the model before I print it.

Oh, and one last thing. When using 123D Catch to create a 3D model, the program does not know anything about size or dimensions. You need to add some reference marks and a reference dimension to your model. One reference dimension will do. In my case, I used the length of one of the vertical side extrusions, which was 24 millimeters. The software figured out all the rest. Do this reference dimension before exporting your object file to Meshmixer.

You might enjoy reading about Mapping with Drones. This article describes how to program your drone quad copter to take pictures of a building, and then create a model of that building using 123D Catch. Enjoy.