Using a depth sensing camera like Kinect, I would like to retrieve the position of an predetermined object (e.g. a cup, fork etc so that I would ultimately be able to grab the object). What would be a way to achieve this?
-
1$\begingroup$ In its current form this question is very complicated. If you could offer some additional information you may get more traction. Specifically, are we to assume that you already know how to how to identify the cup in the point cloud returned by the Kinect? Similarly, do you already know how to generate the grasp required once you know the coordinates of the object? $\endgroup$– DaemonMakerJan 28, 2013 at 23:16
-
$\begingroup$ I've tried to narrow the scope. I would first like to learn to identify an object in space (e.g. by analyzing the point cloud from the Kinect) $\endgroup$– JonJan 29, 2013 at 0:49
2 Answers
If you want to recognise a predetermined object (so one that you have seen before) in a 3d scene. It seems this would work similar to the 2d case: using features. Only in this case using 3d feature extractors/descriptors. You provide your model (so the object that you have seen before), and find salient features in the model. In your new scene you also perform feature extraction and then compare the features with your model. If they match well enough, you have found your object. Have a look at the pcl tutorials on the subject.
You probably want to do this, check it out. They also have code http://personalrobotics.ri.cmu.edu/projects/moped.php
-
$\begingroup$ It helps to add as much content as you can here so that if the content on the other end of the link changes, it is still relevant. $\endgroup$ Feb 17, 2013 at 18:44
-
$\begingroup$ Unfortunately, the topic is not simple, and remains an interesting research problem. The link above summarizes everything quite well. The moped framework is also PhD thesis that you can find here ri.cmu.edu/publication_view.html?pub_id=7326 (the content on this link will not change) $\endgroup$ Feb 17, 2013 at 20:10