[Data Quality] Too noisy lane2d and lane3d

[z-bingo]

I downloaded the OpenLane dataset via your released python client and visualized several cases, noticeable noise labels are found for almost all images, take the below figures as example (red and yellow lines indicate lane2d and lane3d), lane2d and lane3d are just figure to the image by coordinate mapping but no other filters. looking forward to your reply, thx

[ChonghaoSima]

lane3d (shown in yellow) should be visualized in 3D coordinates.

[RicardLee]

Hi, the method of generation of lane3D depends on the accuracy of lidar and sensors calibration. The noise of waymo data and some post process, such as interpolation, may cause this problem. For actual use, We think that fitting the real lane is more important than getting a smooth lane. So we keep some noise for accuracy of lane3D.

[z-bingo]

lane3d (shown in yellow) should be visualized in 3D coordinates.

actually, lane3d is visualized in the 3d coordinates by ignoring the z-axis

[z-bingo]

Hi, the method of generation of lane3D depends on the accuracy of lidar and sensors calibration. The noise of waymo data and some post process, such as interpolation, may cause this problem. For actual use, We think that fitting the real lane is more important than getting a smooth lane. So we keep some noise for accuracy of lane3D.

yep, it works for the actual use, thx