ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange
Ask Your Question

RGBDSLAM -- points of interest

asked 2016-01-11 06:59:56 -0600

goupil35000 gravatar image

updated 2016-01-11 08:42:02 -0600


I need to get all of the points of interest detected with rgbdslam including depth.

So what are the topics to use with RGBDSLAM to get all of the points of interest (SIFT) detected during the creation of the 3D map ?

I think all of these topics will help me: /rgbdslam/online_clouds /rgbdslam/pose_graph_markers

but I'm not sure what are published in these topics. Are they SIFT points (and where is the origin of x, y and z in this case ?). Thanks for reply

edit retag flag offensive close merge delete

1 Answer

Sort by ยป oldest newest most voted

answered 2016-01-13 04:33:49 -0600

updated 2016-01-20 09:52:46 -0600

Rgbdslam doesn't send the features. The topic you mention is for visualization of the poses only.

I've recently wrote a addition to send ORB features on a custom topic, but haven't published it, because I thought there wouldn't be any demand. If you have some ROS and C++ skill, I can create a branch for you that should be easily adapted to SIFT.


Checkout the new branch feature-msg-output on github. It works for sending out ORB features on the topic /rgbdslam/features. The message msg/Feature.msg defines a uint8[] descriptor. For SIFT (or SURF) you probably want to change that to float32[] descriptor

In node.cpp, method Node::toFeatureMessage() you need to reflect that change, by replacing unsigned char with float. And replace the CV_8UC1 in the if-condition by CV_32FC1.

edit flag offensive delete link more


Hello Felix, Thanks for reply.

Yes I have ROS and C++ skills. Tell me where to get this new branch, I will test it and give you feedbacks.

Thanks again !

goupil35000 gravatar image goupil35000  ( 2016-01-15 12:10:06 -0600 )edit

Question Tools

1 follower


Asked: 2016-01-11 06:59:56 -0600

Seen: 172 times

Last updated: Jan 20 '16