fix: OctoMap and Filtered_Cloud Not Updating During Movement Execution

[MarcoMagriDev]

Description

MutuallyExclusive callbackgroup has been added to point_cloud_subscriber_ to allow updating octomap during robot motion. See #2192

---

Additional notes

• I was not able to do the same fix to the depth_image_octomap_updater since image_transport::create_camera_subscription seems to not allow passing SubscriptionOptions.

• I noticed that there is a big computational time mismatch between the two approaches (depth/pointcloud). After digging in a bit i found that the main computational time difference comes from lazy_free_space_updater but unfortunately I have not time to fix it. I will use the pointcloud instead._

[codecov-commenter]

⚠️ Please install the to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

Attention: Patch coverage is 0% with 6 lines in your changes missing coverage. Please review.

Project coverage is 45.61%. Comparing base (75286c7) to head (ce08669).
Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
...octomap_updater/src/pointcloud_octomap_updater.cpp	0.00%	6 Missing ⚠️

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3209      +/-   ##
==========================================
- Coverage   46.01%   45.61%   -0.40%     
==========================================
  Files         714      714              
  Lines       62304    62288      -16     
  Branches     7532     7529       -3     
==========================================
- Hits        28666    28408     -258     
- Misses      33471    33713     +242     
  Partials      167      167              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[sea-bass]

Thanks! To clarify, the fix is simply using a new callback group to process these subscriber callbacks separately from everything else, but the type of callback is less relevant in this case?

Anyways, just one very small comment for readability.

[sea-bass]

Also seems you are right about the image_transport CameraSubscriber not having an input for rclcpp::SubscriberOptions.

I looked at https://github.com/ros-perception/image_common and it wouldn't be too hard to add by just copying how it's done for the Subscriber... but it is extra work.

[sea-bass]

Great, thank you!

I'm gonna request one more review on this one just in case.

[henningkayser]

Thank you! Do you think we should be switching to another API for depth image processing?

[sjahr]

Thanks! Do you mind opening and issue for

I noticed that there is a big computational time mismatch between the two approaches (depth/pointcloud). After digging in a bit i found that the main computational time difference comes from lazy_free_space_updater but unfortunately I have not time to fix it. I will use the pointcloud instead._

[MarcoMagriDev]

Ofc @sjahr , I just open a new issue here #3216