Revision history [back]

The most important difference between Kinect for XBOX One and Kinect for Windows 2 is the USB3 plug. I bought my Kinect for XBox One for just 30€. I needed one afternoon to mount a USB3 connector (and another connector for 12V), so I saved 170€.

But I think you should not use either of the two. The first problem will be to handle the grater amount of data (USB3!), than the power supply (1.18A@12.2V). This can bee solved.

But the new Kinect suffers from noise and reflections. See my post at PCL-users for more details: http://www.pcl-users.org/Obtaining-a-point-cloud-with-Kinec1-vs-Kinect2-tp4037069p4037508.html

For robot navigation I'm using still the old Kinect. It is very challenging to solve this problems (An experienced researcher can spend months and more.) and the bad resolution of the old Kinect is still enough.

But I think you should not use either of the two. The first problem will be to handle the grater amount of data (USB3!), than the power supply (1.18A@12.2V). This can bee solved.

But the new Kinect suffers from noise and reflections. See my post at PCL-users for more details: http://www.pcl-users.org/Obtaining-a-point-cloud-with-Kinec1-vs-Kinect2-tp4037069p4037508.html

Handling the noise is possible. A simple solution is to do not move for 1sec, take 30 images and compute the average. Reflections can be maybe "solved" by restrictions to the environment (and the materials). But for example in the laboratory in the university the new Kinect is entirely useless for navigation at the moment, because of the flooring, the wall paint and some furnishing.