0
$\begingroup$

Rosanswers logo

I built a octomap using rgbdslam package from ros. Now, I have a built map. It consists only the depth information or pose of the objects in 3D. If I use kinect for navigation it will not able avoid the object within distance 1.2m (minimum range of kinect depth sensor).

  1. Is there any available alternatives for the depth sensor to avoid object with in 1.2m?
  2. Or any alternatives (like other type of map from same algorithm or any)?

Thank you in advance for the suggestions.


Originally posted by Sudhan on ROS Answers with karma: 171 on 2012-07-23

Post score: 0


Original comments

Comment by Felix Endres on 2012-07-23:
Sorry, your question is unclear to me. If you can localize in the recorded map, the map will tell you about the obstacles (if they are in it).

Comment by Sudhan on 2012-07-23:
But what about moving objects come in between 1.2m?

$\endgroup$

2 Answers 2

0
$\begingroup$

Rosanswers logo

  1. Yes, mainly laser scanners. Stereo vision or any other sensor with a better min-range will work.

  2. Probably no, if the sensor doesn't give you any data an algorithm cannot do anything.

I said probably because, if you don't get max-range readings in the kinect data if the obstacles are nearer than min-dist, then the octomap doesn't map free space. So the cells where there is no data will be unknown. Your algorithms can now treat unknown cells as obstacles or more precisely you should only plan through free space.

There are some caveats however: Noisy distant readings might still get you free space. Also, you'd probably have to get free space cells for the full height of the robot now and thus might need to actively explore the 3D space.


Originally posted by dornhege with karma: 31395 on 2012-07-23

This answer was ACCEPTED on the original site

Post score: 2

$\endgroup$
0
$\begingroup$

Rosanswers logo

You can use the nyco zoom lens for the kinect. It will lower the minimum distance. Of course the depth measurements will be distorted, but just for obstacle avoidance it should be usable.


Originally posted by Felix Endres with karma: 6468 on 2012-07-24

This answer was NOT ACCEPTED on the original site

Post score: 1


Original comments

Comment by Sudhan on 2012-07-24:
thanks for the suggestion

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.