Conversation
|
@jkleiber @linglejack06 should a filter by target distance be added? this would require writing our own pose estimator class. |
aidnem
left a comment
There was a problem hiding this comment.
Other than the absolute values for the pitch suggestion, this looks fine to me.
Could we also add roll filtering?
|
|
||
| for (PhotonTrackedTarget target : targets) { | ||
| if (target.getPoseAmbiguity() < VisionConstants.maximumAmbiguity | ||
| && target.getPitch() < VisionConstants.maximumPitch) { |
There was a problem hiding this comment.
Can we use an absolute value so that if pitch is too positive or too negative it'll reject both?
There was a problem hiding this comment.
Also is there any reason to do this here and not in the filterPhotonPose method (or we could do it in both places)?
I'm not necessarily asking for you to change it, I'm just curious about the reasoning behind putting it here vs. somewhere else.
|
I added adjustments to variance calculating- could you see if it actually makes any logical sense? thank you! |
|
Should errors based off of distance be calculated based on distance alone or x and y components of distance? |
9438495 to
f37fc52
Compare
…pt at a tunable variance calculator that factors in differences in distance/heading/ambiguity
f37fc52 to
10699b3
Compare
What else should I add? Created a method for filtering individual targets, unsure if it works better