-
Notifications
You must be signed in to change notification settings - Fork 16
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
fix: add downsample_after concat #175
Conversation
Signed-off-by: badai-nguyen <[email protected]>
6219a82
to
fe1358f
Compare
Signed-off-by: badai-nguyen <[email protected]>
@@ -33,7 +33,7 @@ def launch_setup(context, *args, **kwargs): | |||
name="concatenate_data", | |||
remappings=[ | |||
("~/input/twist", "/sensing/vehicle_velocity_converter/twist_with_covariance"), | |||
("output", "concatenated/pointcloud"), | |||
("output", "concatenated/pointcloud_raw"), |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
concatenated/pointcloud_raw
sounds very weird.
The topic name should represent its status, I think. So, it should be something like concatenated/pointcloud
-> concatenated/downsampled/pointcloud
.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@miursh thank you for your comment. My intention is want to keep using concatenated/pointcloud
input for perception including centerpoint, ground_segmentation, etc.
Currently we have 2 options:
- Keep output of pointcloud_preprocessor as
concatenated/pointcloud
and changeconcatenated/pointcloud_raw
-> some more understandable name. For exampleconcatenated/before_downsampled/pointcloud
? - Change output of
pointcloud_preprocessor
and all inputperception
module toconcatenated/downsampled/pointcloud
I currently prefer to 1) but of course I can change as 2) as your opinion.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I prefer another option.
Since I don't want to change the interface between sensing and perception, the output of sensing module should be concatenated/pointcloud
.
And if you want to apply donwsampling before perception, you should add a donwsampling node in /perception
namespace.
BTW, if you change the input point of centerpoint, I want to know that this change does not have much effects on the perception result.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@miursh Thank you for your comment, Since I am going to add the downsampling before common_ground_filter
in perception
(tier4/autoware.universe@bc48f9e) , it will not affect on centerpoint input and this PR is unnecessary. I will close this.
This PR will move voxel_grid_downsample pointcloud from before compare_map_filter to before ground_segmentation to reduce computational cost for X2. This related to tier4/autoware.universe#870