Check if we should be able to see the ball #24
Conversation
| @@ -28,7 +29,6 @@ group_ROS.add("ball_filter_reset_service_name", str_t, 0, "Name of service for r | |||
|
|
|||
|
|
|||
| group_filter.add("filter_rate", int_t, 0, "Filtering rate in Hz", min=0, max=255) | |||
| group_filter.add("min_ball_confidence", double_t, 0, "Minimum ball confidence", min=0.0, max=1.0) | |||
There was a problem hiding this comment.
Why do we not want to filter by confidence?
There was a problem hiding this comment.
I really don't like that this is in here.
There was a problem hiding this comment.
I really do like it, because the approach taken in this pr dosnt work other wise, because of the return in case of a low conficene.
There was a problem hiding this comment.
Also the ball confidence is thresholded 3(!) times in the pipeline with different values. The 2 in the vision have a reason, but this one here is totally unessesary imo.
There was a problem hiding this comment.
AH! then we should meve that to a later point or handle the situation differently. However, I truly believe that the ball filter should be the component selecting the used ball measurements and not the vision or the transformer.
There was a problem hiding this comment.
Lets talk in person about this. I am tried of writing stuff ^^
There was a problem hiding this comment.
What was the result of this discussion?
|
I assume this needs bit-bots/bitbots_vision#290? |
|
I thought about bumping the covariance of the filter, but i didnt find a nice way of doin it. Maybe you arehav an idea |
|
What is the status of this PR? Are the TODOs in the top still TODOs? |
^ |
|
I think this PR is still relevant but not a very high prioetry compared to the ROS2 migration. Maybe @val-ba can have a look at it some time in the future. |
Proposed changes
TODO