tsengapola/costmap_depth_camera

planner_server ObservationDepth size error

mustafaktaasbirfen opened this issue · 6 comments

mustafaktaasbirfen commented

Hello. Have you encountered such an error before? I get my topic from my d435i camera and when I start nav2 I get an error like this.

[planner_server-6] [ERROR] [1703502100.699044300] [global_costmap.global_costmap]: Raw cloud size 150497 is larger than 20000 points. Exiting..
[planner_server-6] [ERROR] [1703502101.010950008] [global_costmap.global_costmap]: Raw cloud size 150168 is larger than 20000 points. Exiting..
[ERROR] [planner_server-6]: process has died [pid 16063, exit code -11, cmd '/home/mustafa/ros2_ws/install/nav2_planner/lib/nav2_planner/planner_server --ros-args --log-level info --ros-args -r __node:=planner_server --params-file /tmp/tmp97b_ghha -r /tf:=tf -r /tf_static:=tf_static -r cmd_vel:=/cmd_vel_unstamped'].

tsengapola commented

The issue is that there are too many points in your topic, you should voxelize them.
I push a new commit which by default enable the voxelization on the input cloud to save the computation.
c101f335143da6313554d9f4ae1dd5986c6f7b98
However, I recommend to voxelize the point cloud outside this plugin, because voxelization might increase costmap_2d loop time.

  • 👍1
mustafaktaasbirfen commented

Thanks. Now I'm struggling with a new error.

[planner_server-6] [ERROR] [1703511718.141410144] [global_costmap.global_costmap]: ObservationDepth size 86424 is larger than 5000 points. Exiting..
[planner_server-6] [ERROR] [1703511719.162417849] [global_costmap.global_costmap]: ObservationDepth size 86966 is larger than 5000 points. Exiting..

tsengapola commented

Thanks for testing.
Try new commit and let me know the result

  • 👍1
mustafaktaasbirfen commented

Hi, thank you. It doesn't give the error anymore, but my pointcloud2 data is not coming in properly. It freezes a lot. That's why I can't add it to the global costmap. I will give you more detailed feedback tomorrow. Good work and thank you again.

mustafaktaasbirfen commented

Hi, good work. How do I reduce the point cloud data in the depth of my D435i camera? I am experiencing freezing and lag when opening rviz because too much data is published.

tsengapola commented

Hello, to reduce the point cloud, you should refer to the voxelize point cloud using Point Cloud Library (PCL).
Downsampling a PointCloud
I suggest you to write a node that subscribe the point cloud and down sample it then publish it.