By: Rob Adams on April 4, 2013
This week has not been a success for me. I thought that it would be fairly easy as I have been doing 3D printing for a while, but it turns out that 3d scanning has been a challenge. Together with illness and technology problems, I am nowhere near where I had hoped my 3d Scanning capabilities would be. I will outline my problems, and then conclude with my (limited) successes. I was all prepared to set up the Kinect scanner, but have not yet managed to do so. First I ran into problems with computers and system requirements. When I finally switched to another, newer computer. I downloaded ReconstructMe, with only a few glitches. That installation took about 20 minutes. Next I opened the software and connected the Kinect. It downloaded the drivers, and began installation, but failed at the point of the XBox audio point. I could not get the system to open again to the installation menu and it continued to close with system error messages. I went to the old stand-by remedies of restarting the computer and uninstalling/reinstalling the software. And the installation failed at that point. I did some internet searching, but could not find a solution, so I sent off an email for assistance. In the meanwhile, I decided to try 123D Catch. I was excited about this method of scanning, as I travel to archaeological sites, many of which are undocumented. I have hundreds of pictures of a former palace in Samarqand, Uzbekistan where I did my doctoral thesis research. Having read through the tips in the Getting Started with MakerBot section on 3D scanning, I decided to use as my test subject a troll doll from my mother’s childhood (who now resides with me, so I thought a 3D print of the doll for my mother would be a great gift).
I dutifully placed the object on a surface where I could obtain photos for a set distance at all angles, and with objects in the background. After uploading the photos, I was excited to see the result … which was a troll with his front and backside overlapping. I tried remapping, by matching points, but was unable to create a decent 3D image with a front AND backside. I also tried taking fewer photos, but that produced a similar result. Next I switched to a Coke bottle, with again no success. I suspect I am not picking good objects. My initial thought was that both the troll and the Coke bottle were clear with distinctive points, but maybe they coloring on each object is too uniform. I watch videos on ReconstructMe and NetFabb, but I was unable to download NetFabb due to repeated error messages.
After a day and a half, and running out of time, I am posting this but will continue working on the 3D scanning, and will add results when achieved. I would really like to get the Kinect working and better understand 123D Catch. Once I am able to do both, I want to do a series of tests with various objects and compare to two methods of 3D scanning. I did have time to print several things while waiting for files to render and software to download. I used several projects from Thingiverse, even customizing a few. My only problem with printing, aside from wanting a larger print area, is removing the raft and fill.