NTU数据集SLAM运行失败

[Stephen1e]

十分感谢您的工作！这对我非常有帮助！
我尝试在NTU数据集（https://ntu-aris.github.io/ntu_viral_dataset/）
上运行您的改进代码，但得到的地图是z轴颠倒的，算法估计轨迹的漂移也很大，下面是我的param列表
#camera

image_width: 752
image_height: 480
distortion_parameters:
k1: -0.300267420221178
k2: 0.090544063693053
p1: 3.330220891093334e-05
p2: 8.989607188457415e-05
projection_parameters:
fx: 4.313364265799752e+02
fy: 4.327527965378035e+02
cx: 3.548956286992647e+02
cy: 2.325508916495161e+02

acc_n: 6.0e-2 # accelerometer measurement noise standard deviation.
gyr_n: 5.0e-3 # gyroscope measurement noise standard deviation.
acc_w: 8.0e-5 # accelerometer bias random work noise standard deviation.
gyr_w: 3.0e-6 # gyroscope bias random work noise standard deviation.
g_norm: 9.805

extrinsicRotation: !!opencv-matrix
rows: 3
cols: 3
dt: d
data: [-0.01916508, -0.01496218, 0.99970437,
0.99974371, 0.01176483, 0.01934191,
-0.01205075, 0.99981884, 0.01473287]
#Translation from camera frame to imu frame, imu^T_cam
extrinsicTranslation: !!opencv-matrix
rows: 3
cols: 1
dt: d
data: [0.00519443, 0.1347802, 0.01465067]

#lidar
extrinsicTranslation: [0.05, 0.0, -0.055]
extrinsicRotation: [ 1, 0, 0,
0, -1, 0,
0, 0, -1]

[Cc19245]

十分感谢您的工作！这对我非常有帮助！ 我尝试在NTU数据集（https://ntu-aris.github.io/ntu_viral_dataset/） 上运行您的改进代码，但得到的地图是z轴颠倒的，算法估计轨迹的漂移也很大，下面是我的param列表 #camera

image_width: 752 image_height: 480 distortion_parameters: k1: -0.300267420221178 k2: 0.090544063693053 p1: 3.330220891093334e-05 p2: 8.989607188457415e-05 projection_parameters: fx: 4.313364265799752e+02 fy: 4.327527965378035e+02 cx: 3.548956286992647e+02 cy: 2.325508916495161e+02

acc_n: 6.0e-2 # accelerometer measurement noise standard deviation. gyr_n: 5.0e-3 # gyroscope measurement noise standard deviation. acc_w: 8.0e-5 # accelerometer bias random work noise standard deviation. gyr_w: 3.0e-6 # gyroscope bias random work noise standard deviation. g_norm: 9.805

extrinsicRotation: !!opencv-matrix rows: 3 cols: 3 dt: d data: [-0.01916508, -0.01496218, 0.99970437, 0.99974371, 0.01176483, 0.01934191, -0.01205075, 0.99981884, 0.01473287] #Translation from camera frame to imu frame, imu^T_cam extrinsicTranslation: !!opencv-matrix rows: 3 cols: 1 dt: d data: [0.00519443, 0.1347802, 0.01465067]

#lidar extrinsicTranslation: [0.05, 0.0, -0.055] extrinsicRotation: [ 1, 0, 0, 0, -1, 0, 0, 0, -1]

你好，我之前就注意到了这个问题。NTU数据集的作者给出了适配LIO-SAM的代码，他对代码进行了一些修改，我之前给他提过 issue 询问，你可以参考一下。

[sp225]

十分感谢您的工作！这对我非常有帮助！ 我尝试在NTU数据集（https://ntu-aris.github.io/ntu_viral_dataset/） 上运行您的改进代码，但得到的地图是z轴颠倒的，算法估计轨迹的漂移也很大，下面是我的param列表 #camera

image_width: 752 image_height: 480 distortion_parameters: k1: -0.300267420221178 k2: 0.090544063693053 p1: 3.330220891093334e-05 p2: 8.989607188457415e-05 projection_parameters: fx: 4.313364265799752e+02 fy: 4.327527965378035e+02 cx: 3.548956286992647e+02 cy: 2.325508916495161e+02

acc_n: 6.0e-2 # accelerometer measurement noise standard deviation. gyr_n: 5.0e-3 # gyroscope measurement noise standard deviation. acc_w: 8.0e-5 # accelerometer bias random work noise standard deviation. gyr_w: 3.0e-6 # gyroscope bias random work noise standard deviation. g_norm: 9.805

extrinsicRotation: !!opencv-matrix rows: 3 cols: 3 dt: d data: [-0.01916508, -0.01496218, 0.99970437, 0.99974371, 0.01176483, 0.01934191, -0.01205075, 0.99981884, 0.01473287] #Translation from camera frame to imu frame, imu^T_cam extrinsicTranslation: !!opencv-matrix rows: 3 cols: 1 dt: d data: [0.00519443, 0.1347802, 0.01465067]

#lidar extrinsicTranslation: [0.05, 0.0, -0.055] extrinsicRotation: [ 1, 0, 0, 0, -1, 0, 0, 0, -1]
请问这个问题解决了麻，我也遇到了同样的问题，可不可以分享一下解决方案呢，十分感谢～