ROS Answers SE migration: Kinect and PCL

Question

Hey everyone

I am currently working on a project where I have a pile of cloth and kinect camera mounted above this pile looking down on it. The aim is then to find a grasping point using the kinect camera so a robot arm is able to grasp a piece of cloth. My initial idea was to use the depth information to find the deviation between different pixels, an area with high deviation will then correspond to a piece of cloth sticking out. A good example could be this image where the folds will give a high deviation and thus a good grasping point. A thing I have noticed during my research is the widely use of pointclouds and the PCL library with the Kinect camera, and therefore I wonder about using the pointclouds to find the deviation. My problem is that I cant find a good indeep tutorial how to use the PCL library and I wonder if someone has some good links or explanation how to process pointcloids with the PCL library.

---

Originally posted by Aslund on ROS Answers with karma: 289 on 2011-04-05

Post score: 0

Answer 1

The tutorials should get you started. Look at the openni stack how to get your kinect to publish pointclouds. Also notice that PCL has it's own mailing list on which you can post questions.

---

Originally posted by KoenBuys with karma: 2314 on 2011-04-05

This answer was ACCEPTED on the original site

Post score: 1

---

Original comments

Comment by Aslund on 2011-04-05:
Hey, thanks for your answer. I have previously looked at the tutorials on the PCL homepage, but their look very concrete and specific in my eyes and I had hoped there was some more basic guides that more focuses on the fundamentals and maybe short introduce the core methods in the library