Some issue about reprojection error and imu error

[Richardcvpr]

The calibtation results is not good enough. I think there may be some issues with IMU, but I don't know where the problem is.

Calibration results
===================
Normalized Residuals
----------------------------
Reprojection error (cam0): mean 4.389178270980999, median 3.4851612083506827, std: 3.546836339123121
Gyroscope error (imu0):
mean 4.987582349959954, median 2.840098285394952, std: 5.739188841909919
Accelerometer error (imu0): mean 5.648619958147809, median 4.5686257632248255, std: 3.9259018299971866
Residuals
----------------------------
Reprojection error (cam0) [px]: mean 4.389178270980999, median 3.4851612083506827, std: 3.546836339123121
Gyroscope error (imu0) [rad/s]: mean 0.04001092025534412, median 0.02278357288981692, std: 0.046040388102236
Accelerometer error (imu0) [m/s^2]: mean 0.7112548280396047, median 0.5752656676632362, std:
0.49433607707441224
Transformation (cam0):
-----------------------
T_ci: (imu0 to cam0):
[[ 0.92757641 -0.27752758 0.25016086 -0.05011873]
[ 0.30536953 0.94890312 -0.07957594 0.01329463]
[-0.2152939 0.15020427 0.96492861 0.33276062]
[ 0.0.0.1.]]
T_ic: (cam0 to imu0):
[[ 0.92757641 0.30536953 -0.2152939 0.11407051]
[-0.27752758 0.94890312 0.15020427 -0.07650671]
[ 0.25016086 -0.07957594 0.96492861 -0.30749456]
[ 0.0.0.1.]]
timeshift cam0 to imu0: [s] (t_imu = t_cam + shift)
-7.978506571910605
Gravity vector in target coords: [m/s^2]
[-0.1185964 0.31407217 9.80080183]Calibration configuration
=========================
cam0
-----
Camera model: pinhole
Focal length: [287.5931203196147, 318.44992107782025]
Principal point: [381.33348613478404, 245.2892815651306]
Distortion model: radtan
Distortion coefficients: [0.014059, -0.021049, 0.001291, 0.003152]
Type: aprilgrid
Tags:
Rows: 6
Cols: 6
Size: 0.021 [m]
Spacing 0.005999999999994 [m]
IMU configuration
=================
IMU0:
----------------------------
Model: calibrated
Update rate: 100.0
Accelerometer:
Noise density: 0.01259165660479
Noise density (discrete): 0.1259165660479
Random walk: 0.000426938443672722
Gyroscope:
Noise density: 0.000802210719501511
Noise density (discrete): 0.00802210719501511
Random walk: 2.82325427421011e-05T_ib (imu0 to imu0)
[[1. 0. 0. 0.]
[0. 1. 0. 0.]
[0. 0. 1. 0.]
[0. 0. 0. 1.]]
time offset with respect to IMU0: 0.0 [s]

imu_image_good_2024-06-03-17-58-28-report-imucam.pdf

Can anyone explain me, what's wrong with calibration results? I would be thankful for any answer. Thanks