Closed ChengWang2018 closed 1 year ago
@ChengWang2018 I think this depends on how you transformed the lanes in the map from the global coordinate system to the pixel coordinate system in the first place
I have not done that. I use images for the lane detection. Thus, the mp is firstly not needed. So, I donot quite understand your ideas. Could you explain it a bit more? Thanks!
There's a method map_pointcloud_to_image
which maps a set of 3D coordinates to a given image - perhaps you can use some of the ideas in there to do the reverse
Hi,
I use a lane detection algorithm to detect the lanes in Nuscene dataset. The results (points) are on the pixel coordinate system, How can I convert it global coordinate system so that I can compare the results with lanes in the map.
Thanks a lot!