Range scan captured using KinFu.
Today I managed to thoroughly test the Kinect and everything seems OK.
I couldn’t resist setting up a project and getting to know the libraries and the all-encompassing mess of inclusions it requires. After the customary headaches, I managed to get the PCL rolling and capturing a simple point cloud as you can see in the image caption (my first capture!). I also set-up a SVN repository for the coming weeks of work. Now I need to get used to a style of programming I’m not really comfortable with.
Today was an important day as I received some very helpful advice to get my work with 3D capture underway. This came from Bruno de Araujo, a PhD student at Instituto Superior Técnico (check out his page). He lent me some of his expertise around capture with Microsoft Kinect, the state of the art, about some of top names also working with 3D capture and about the device itself. I’ll write more about that soon as I follow up those threads.
Also today I received the Microsoft Kinect I will be working during this process from my research mentor, and didn’t lose much time following some of the advice I got from Bruno. This involved exploring the Point Cloud Library SDK which I promptly downloaded and proceeded to build along with its dependencies. The process took about 5 hours of grueling memory consumption and heavy processing during which I took upon myself to update the blog. Rarely have I seen my laptop struggling so much with anything. Auspicious start it is!
I expect tomorrow to have everything working so I can run some tests on the Kinect.
Also I added a couple links to the blogroll on the right. One of them is the 3DORuS program where my project is inserted. The other is a similar blog from a fellow colleague which will be working on Model Retrieval for the LTouchIt project.