Attention: Answers.ros.org is deprecated as of August the 11th, 2023. Please visit robotics.stackexchange.com to ask a new question. This site will remain online in read-only mode during the transition and into the foreseeable future. Selected questions and answers have been migrated, and redirects have been put in place to direct users to the corresponding questions on Robotics Stack Exchange. Additional details are available here.

optimal kinect/xtion placement with regard to object recognition

asked 2012-04-01 19:36:02 -0500

bit-pirate

2062 ●24 ●47 ●55

updated 2016-10-24 09:03:02 -0500

ngrennan
1 ●1 ●1

Hi!

I am wondering, if there is some kind of best practice out there or someone could share her/his experience about how to best place the Kinect, Xtion or a similar device on a robot.

To be more specific, I am thinking about which viewing angle is necessary to get good results for object recognition, e.g. with objects on a table. Is there an optimal viewing angle - maybe 45 degrees?

I would be happy, if somebody doing table-top manipulation, e.g. with the PR2 + Kinect, could tell me about the viewing angles for which object recognition works well.

Thank you!

edit retag flag offensive close merge delete

add a comment

1 Answer

Sort by » oldest newest most voted

answered 2012-07-23 11:13:47 -0500

Vincent Rabaud

1111 ●25 ●20

well, it all depends on the technique you use, it is that simple: template techniques like LINEMOD will need the same view points that you acquired during training, feature based ones can be a bit more robust to warping and so on.

Train your detector in angle/scale/illumination in the same way you will detect your object and accuracy will increase.

edit flag offensive delete link

add a comment

Question Tools

1 follower

Stats

Asked: 2012-04-01 19:36:02 -0500

Seen: 695 times

Last updated: Jul 23 '12

optimal kinect/xtion placement with regard to object recognition edit

1 Answer

Question Tools

Stats

Related questions

optimal kinect/xtion placement with regard to object recognition