There is a bit of confusion on the configuration of VINS-Fusion. I am testing both VINS-Mono and VINS-Fusion with CARLA Simulation Environment. VINS-Mono is working well with the following transform:
#Rotation from camera frame to imu frame, imu^R_cam
# extrinsicRotation is taken from the API output of the MYNTEYE camera
extrinsicRotation: !!opencv-matrix
rows: 3
cols: 3
dt: d
data: [ 0.0, 0.0, 1.0,
-1.0, 0.0, 0.0,
0.0,-1.0, 0.0]
#Translation from camera frame to imu frame, imu^T_cam
extrinsicTranslation: !!opencv-matrix
rows: 3
cols: 1
dt: d
data: [0.06,0.0, 0.0]
For VINS-Fusion the similar config is not working:
These transforms are from the Camera frame to IMU frame ? In case VINS-Mono it seems to be so. However I doubt about it regarding VINS-Fusion. Because when I see kitti_config00-02.yaml, I see that:
Oh man, I had made a terrible mistake with the distance between the stereo cameras. Thats why VINS-Fusion was failing!!!
Using m instead of cm between the stereo camers!
Hello @shaozu / @shaojie
There is a bit of confusion on the configuration of VINS-Fusion. I am testing both VINS-Mono and VINS-Fusion with CARLA Simulation Environment. VINS-Mono is working well with the following transform:
For VINS-Fusion the similar config is not working:
These transforms are from the Camera frame to IMU frame ? In case VINS-Mono it seems to be so. However I doubt about it regarding VINS-Fusion. Because when I see kitti_config00-02.yaml, I see that:
Now I know in KITTI Vehicle setup this is not the Camera to IMU transform. So I think clarification and better documentation is needed.