[axoid]

I've just started playing with a new technology and I wanted to share and get some thoughts and ideas.

I hooked a Kinect sensor to a PC and started 3D scanning objects, namely my Camaro. I'm thinking that this could be useful when it comes to further modifying the car.

I've been wanting to flair the fenders and if I can get a good model of the fender I could then model a flare to it and then 3D print the flair and then just flip the model and print it again for the other side, and do it as many time as I want.

Scanning parts could also lead to engineering better replacements parts in CAD software in the future. Anything from front 3rd gen spindles to a IRS for 3rd and 4th gens. I've seen articles about using 3D printing to create molds for casting parts or of course models for CNC machines.

This is one of my first test scans that I did of the engine bay.
(IMG:http://www.incom-racing.com/wp/wp-content/uploads/2014/04/engine-test.jpg)

I need to work on the software's settings and I used a i3 laptop to do the processing. I think things would go smoother with more horsepower in the PC. The software uses the graphics chip to do some of the heavy lifting.

I looking at getting a 3d printer this summer, though it won't be big enough to print a fender, but it is a start. I would love to build a 40" Reprap printer.

Keep in mind, I'm an IT guy, not a mechanical engineer.

[JimMueller]

I'm an IT fella also... I've got an Xbox360 and bought my Kinect sensor separately. I also have 1) a 8GB i5-2520M laptop and 2) a 16GB i7-3770K desktop that I'd be willing to play on. Is there a way to separate the scanning and processing functions, so perhaps I could scan with the slower mobile i5 and process on the faster i7? Or someone could scan remotely and send the files to someone else to process faster?

Can you provide links to all the software/tools?

[axoid]

I'm not sure what you mean by "processing", I'm guessing that your talking about the actual generation of the mesh model. I think it is possible to generate the mesh later by just recording the depth data and running the integration later, but you would have to write your own software to do it.

The software the Microsoft has out, generates the mesh in real-time buy using the 3D processing built into the graphics card. Also the more move around an area, getting your object from multiple angles, the more it improves the accuracy of the mesh. Microsoft's tool is limited to some extent, but it is just an example in the Software Development Kit. So if you know C++ or C# you can make it do more that the examples are capable of.


The "Kinect Fusion" scanning tool is part of the Kinect SDK (Software Development Kit).
http://www.microsoft.com/en-us/kinectforwi.../Downloads.aspx

This is a detailed article on Kinect Fusion.
http://msdn.microsoft.com/en-us/library/dn188670.aspx


Additional software:

Meshlab is a editing tool that allow you to modify the 3D meshes. It helps when it comes to 3D printing. (open source)
http://meshlab.sourceforge.net/

Blender is a 3D animation tool and has a very good 3D modeling features. I haven't done much with this yet. (open source)
http://www.blender.org/

This is an open source CAD tool, but I haven't tried it.
http://sourceforge.net/projects/free-cad/