The problem with constructing a map with two kinetics 2.0

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

The problem with constructing a map with two kinetics 2.0

baltic
Hello,
I'm trying to run Rtabmap SLAM using two kinetics 2.0. Both sensors and rtab visual odometry work correctly, but instead of a 3D map I get a few incomprehensible stripes of pixels passing through the position of one of the cameras. However, Rtabmap does not report errors.



I'm using the latest version of rtabmap and rtabmap_ros downloaded from github.

Kinect and rtab settings files are in the application
two_kin.launch
2_kinect2_bridge_1.launch
2_kinect2_bridge_2.launch
Reply | Threaded
Open this post in threaded view
|

Re: The problem with constructing a map with two kinetics 2.0

matlabbe
Administrator
Hi,

The blue line strips are laser scan points. "gen_scan" parameter is failing creating a fake laser scan from the kinects. I tried the demo "as is" and it works. In your launch file, you set:
<node pkg="tf" type="static_transform_publisher" name="base_to_camera2_tf"
      args="0.35 0.0 0.0 0.0 1.57 0.0 /base_link /camera2_link 100" />
The 90 deg pitch rotation is the problem. I tried with yaw rotation like the demo "0.35 0.0 0.0 1.57 0.0 0.0" and the laser scans are generated correctly. I don't know exactly how you fixed the kinects, but if they are not placed in a configuration similar as the demo (see screenshot below from this post), set gen_scan to false to avoid creating laser scans. Set Grid/FromDepth to true if you want cloud_map from the kinect clouds.



cheers,
Mathieu