Lidar2imu custom data problem on auto_calib

[CurssedCoffin]

I'm using my own pcd data and the head is like this

# .PCD v0.7 - Point Cloud Data file format
VERSION 0.7
FIELDS x y z intensity
SIZE 4 4 4 1
TYPE F F F U
COUNT 1 1 1 1
WIDTH 1
HEIGHT 66243
VIEWPOINT 0 0 0 1 0 0 0
POINTS 66243
DATA binary

The intensity is uint8_t and I changed the custom pointcloud structure in auto_calib/include/common/Lidar_parser_base.h

struct LidarPointXYZIRT {
  PCL_ADD_POINT4D;
  uint8_t intensity;
  EIGEN_MAKE_ALIGNED_OPERATOR_NEW
} EIGEN_ALIGN16;

POINT_CLOUD_REGISTER_POINT_STRUCT(
    LidarPointXYZIRT,
    (float, x, x)(float, y, y)(float, z, z)(uint8_t, intensity, intensity))

But when i'm running run_lidar2imu I got warnings showing failed to find match for field 'intensity'

And the ceres iteration also shows like

==>ROUND 22
deltaQ: 1 0 0 0
solver.cc:415 Function tolerance reached. No non-constant parameter blocks found.
Ceres Solver Report: Iterations: -2, Initial cost: 0.000000e+00, Final cost: 0.000000e+00, Termination: CONVERGENCE
delta rpy: 0 0 0
delta T: 0 0 0
Round Finish!

How to solve this problem?

Thanks!

[CurssedCoffin]

The pcd data is extracted from rosbag by the way

[CurssedCoffin]

I have solved this problem by debugging.

[luameows]

@CurssedCoffin did u try their auto_calib? What about the result?

[CurssedCoffin]

@CurssedCoffin did u try their auto_calib? What about the result?

The Lidar-IMU auto calib result is not good on my dataset of Robosense RSBP lidar and a GPS+IMU device. The extrinsic matrix is not so accurate. When I use a 1 min data of rosbag, and project lidar frames to one single frame, it's not closed and have some edges on it.