I am exploring the capabilities that a Kinect has, that can be used in the robotics field.
I am working on a telepresence robot where the Kinect is an important component of the robot. It is in fact the component that is used in order to generate a map of where the robot is located, so that it can navigate smoothly and avoid the obstacles in front of it. As I searched on similar projects, I found some applications that are closer, if not the same of what I want to do. The closest application to my research is the folling project done by MIT
What I know so far is that the techniques involved in that process are:
- 3D Mapping
- 3D point Cloud
- Simultaneous Localization and Mapping (SLAM)
I currently succeeded in generating a 3D point cloud using the following tutorial, but did not know what to do next.
I would love to have any information that can be helpful, either a tutorial or a library or something similar. Basically a starting point
I think that this library is what you are looking for:
http://pointclouds.org/
And I'm actually quite surprised that you haven't found it before as you are using the tag of that precise library on your question.
It has a bunch of tutorials here:
http://pointclouds.org/documentation/tutorials/#i-o
Many of those related with the use of Kinect.
And for given you a little more guidance on your personal interest, you should check this implementation of KinFu:
http://pointclouds.org/documentation/tutorials/using_kinfu_large_scale.php#using-kinfu-large-scale
Hope this is what you were looking for.