search

ntu-aris / ntu_viral_dataset

MIT License
138 stars 14 forks source link

Question abourt the vertical LiDAR's extrinsics #2

Closed peterWon closed 3 years ago

peterWon commented 3 years ago

Thanks a lot for your work!

Problem

Recently, I'm tring to test my SLAM method using your multi-lidar dataset. However, I find a problem in the provided calibration parameters. In detail, first, the last item of the vertical lidar's extrinsic should be 1, not 0.

Second, I transform the 'T_Body2Lidar' to roll, pitch and yaw, resulting in 1.57, 3.14, 0 respectively. I orginzed the two lidars and the IMU into one urdf file, and checked the projected clouds in rviz. But the result showing the pointcloud of the vertical lidar is upside down with the provided extrinsics. When I change its euler angles to '1.57, 0, 3.14', the result seems good. I don't know where goes wrong. Could you please give me some cues, thanks!

rviz_screenshot_2021_04_14-15_14_36

Reproduce

Here are related scirpts to reproduce the error. tf_debug.zip

brytsknguyen commented 3 years ago

Hi Peter,

How do you calculate the euler angles?

For the provided extrinsics [-1 0 0; 0 0 1; 0 1 0], the matlab command rotm2eul([-1 0 0; 0 0 1; 0 1 0]) returns the yaw-pitch-roll sequence [3.14, 0, 1.57], which is indeed the result that gives you the correct aligment.

peterWon commented 3 years ago

Thanks for your quick reply! I resorted to Eigen to convert the matrix to euler angles. I'll check my codes again.

peterWon commented 3 years ago

@brytsknguyen After checking my codes, I confused Eigen's eulerAngles order. Thank you again. I'll close this issue.