Assume a team is attacking towards the right. The coloured dots indicate goalposts. If teams are detecting goalposts, it is useful to know whether a goalpost is a left goalpost or a right goalpost.
It seems pretty clear, that red is a left goalpost and blue is a right goalpost. But what about yellow and black, which one is left and right?
If a robot is on the field looking back towards its own goals, it makes sense for the black post to be the left post, and the yellow post to be the right post.
However, if we define x and y coordinates on the field as below, it seems logical to define positive-y as “left” and negative-y as “right”. Then, the yellow post is the left post and the black post is the right post.
I think we at the Bit-Bots only use coordinates for the goalposts and not “left”/“right” if I remember correctly. If we imagine a goal in front of us from the center point and call the left one in our view the left and the right one the right it is quite intuitive, but it kind of breaks when we talk about strategy in a global frame or if the goalie stands in the goal and has the right one on its left. I therefore would suggest skipping “right” and “left” and address them as “positive y” and “negative y” in the global frame. I am fine to descibe them as “left” and “right” if we talk about detections in the image space/frame and not cartesian ones, as it is clear that this is based on the viewpoint of the robot.
@flova Welcome to RoboCup Cross-Cutting — thanks for contributing!
The point you made makes a lot of sense and I agree.
To reiterate, if the image given below was from a robot’s camera, it would make sense to call this the left goal post from the robot’s position, but not in the field frame.
