I’ve been busy experimenting with the Xbox Kinect and exploring some of its capabilities. The Kinect sensor utilises infra-red depth sensors to sense a space or objects in 3D, something I explored in my recent MA project Bound. I haven’t really touched the tip-of-the-iceberg of what this technology enables. It’s completely under-utilised in game consoles, and i’m only using the first-gen Xbox 360 version.
I discovered the Skanect software for Mac/PC. It enables the scanning of objects in 3D space for use in software or to 3D print. There’s two versions, one free feature cut down version and a full version at a reasonable €99. I’ve just explored the free version, the main feature missing is the limited resolution export into 3D software, other than that it mostly works in the same way.
It has various presets to scan either small objects, humans or rooms, although you can customise these settings. It needs a pretty beefy computer to run well, although it chugs along on my 2009 MacBook Pro, using the CPU rather than the GPU to render the scene. I still managed to get some pretty good results. You scan the object by moving the Kinect sensor round it, when you’ve mapped the whole object the data is rendered into a 3D object. The software then computes any gaps in the data and fills any blank holes in the object. The Kinect uses a traditional (lo-res) VGA camera to record images, which it then maps as textures onto the completed object.
My colleague Eric Renno helped scan me in! You can upload your final objects online and send them off for print. With the full version you can open the models in 3D software, which then enables editing options. If anything though, it’s good fun!