run module¶
Run the SLAM system and save, at each frame, the current depth and pose
usage: run [-h] [--dataset DATASET] [--settings SETTINGS] [--dest DEST]
[--pose_id POSE_ID] [--is_evaluate_depth] [--is_evaluate_pose]
[--is_bash] [--data_type {TUM,KITTI_VO,KITTI,OTHERS}]
[--gt_depth GT_DEPTH] [--gt_pose_dir GT_POSE_DIR]
[--gt_pose_txt GT_POSE_TXT] [--align {scale,scale_7dof,7dof,6dof}]
[--named NAMED]
Named Arguments¶
- --dataset
path to dataset
Default: “/media/Datasets/KITTI_VO/dataset/sequences/10”
- --settings
which configuration?
Default: “./settings_kitty.yaml”
- --dest
where do we save artefacts?
Default: “./results_kitty_vo_10”
- --pose_id
between which frames do you want compute the pose? If pose_id==-1, get the pose between 0->T; if pose_id >0, compute the pose between T-pose_id->T For instance, if pose_id=2 then compute the pose between T-2->T
Default: -1
- --is_evaluate_depth
If set, will evalute the orb depth with the gt files
Default: False
- --is_evaluate_pose
If set, will evalute the orb pose with the gt files
Default: False
- --is_bash
If set, means use bash shell to evaluate
Default: False
- --data_type
Possible choices: TUM, KITTI_VO, KITTI, OTHERS
which dataset type
Default: “KITTI_VO”
- --gt_depth
the gt depth files of the dataset
Default: “/media/Datasets/KITTI_VO_SGM/10/depth”
- --gt_pose_dir
each frame’s gt pose file, saved as previous to current, and filename as current.npy
Default: “/media/Datasets/KITTI_VO_SGM/10/npy_pose”
- --gt_pose_txt
this is the gt pose file provided by kitty or tum.
Default: “/media/Datasets/KITTI_VO/dataset/poses/10.txt”
- --align
Possible choices: scale, scale_7dof, 7dof, 6dof
alignment type
Default: “7dof”
- --named
the names for saving pose
Default: “kitty_vo_10”