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

rgbdslam_v2 not receiving any video stream from Realsense R200

asked 2016-04-08 14:50:51 -0600

archit gravatar image

updated 2016-04-08 17:10:26 -0600

I'm trying to use a Realsense R200 camera to generate 3-D maps. For this purpose, I have the realsense_r200_nodelet running on $ROS_MASTER_URI like so

roslaunch realsense_camera realsense_r200_nodelet_standalone_manual.launch color_fps:=30 color_height:=480 color_width:=640 depth_fps:=30 depth_height:=480 depth_width:=640

I have rgbdslam running on a separate machine. I run it like so

roslaunch rgbdslam rgbdslam.launch

Unfortunately, I don't see any video being captured on the GUI that comes up. I have updated the rgbdslam.launch file to have the right values for the config parameters. The relevant ones are as shown below

<param name="config/topic_image_mono" value="/camera/color/image_raw"/>
<param name="config/topic_image_depth" value="/camera/depth/image_raw"/>
<param name="config/topic_points" value="/camera/depth/points"/>
<param name="config/base_frame_name" value="/camera_link"/>
<param name="config/camera_info_topic" value="/camera/color/camera_info"/>

I have confirmed using rviz that my machine does receive video stream from the Realsense camera. Does anyone have any pointers on what I might be doing wrong to not receive anything on the rgbdslam GUI.

edit retag flag offensive close merge delete


Did you check that you are receiving both rgb and depth stream ? Are you sure that your rgb topic is /camera/color/image_raw ? It seems to me that the default was /camera/rgb/image_color.

al-dev gravatar image al-dev  ( 2016-04-08 20:34:15 -0600 )edit

Yeah, I'm receiving both the rgb and depth streams (checked with rviz). I changed the topic to /camera/color/image_raw because that it is the one on which Realsense sends the rgb stream (it works when used with rviz).

archit gravatar image archit  ( 2016-04-09 02:05:38 -0600 )edit

Ok. Which versions of rgbdslam and of the realsense ROS driver are you using ? I would try setting config/topic_points to "". As specified in the launch file : "if empty, poincloud will be reconstructed from image and depth"

al-dev gravatar image al-dev  ( 2016-04-09 22:17:51 -0600 )edit

I'm using rgbdslamv2 (repo) and Intel's package for Realsense (repo). I'm trying what you suggested and it does seem to work. It is just too slow right now.

archit gravatar image archit  ( 2016-04-10 15:01:56 -0600 )edit

The performance can be linked to a number of different factors, you might want to create another ticket for it. If this works I will post it as an answer.

al-dev gravatar image al-dev  ( 2016-04-10 16:39:06 -0600 )edit

2 Answers

Sort by ยป oldest newest most voted

answered 2016-04-10 16:40:19 -0600

al-dev gravatar image

Set config/topic_points to "". As specified in the launch file : "if empty, poincloud will be reconstructed from image and depth"

edit flag offensive delete link more

answered 2016-07-06 09:57:32 -0600

Ziwen Qin gravatar image

I have same question. But I use kinect2, so I installed libfreenect2 and iai_kinect. Using iai_kinect2 bridge translate in kinect2 and ros. But RGBDSLAM GUI seem show point cloud result of fusion of single frame .

edit flag offensive delete link more

Question Tools



Asked: 2016-04-08 14:50:51 -0600

Seen: 986 times

Last updated: Jul 06 '16