Note that we resize the image both in
KITTI dataset and Malaga dataset from original resolution into
Download KITTI_dataset.tar.gz and Malaga_dataset.tar.gz
and extract them into folders KITTI_rec_256 and Malaga_down in $DATA_ROOT. Then download
kitti_ckpt.tar.gz and malaga_ckpt.tar.gz and extract them into
folders kitti_ckpt and malaga_ckpt in $ROOT
We list the sequence index of KITTI Odometry split and corresponding folder name in KITTI raw split (03 is not included in training set as the IMU data is not available):
| Seq index | name |
|---|---|
| 00 | 2011_10_03_drive_0027 |
| 01 | 2011_10_03_drive_0042 |
| 02 | 2011_10_03_drive_0034 |
| 03 | 2011_09_26_drive_0067 |
| 04 | 2011_09_30_drive_0016 |
| 05 | 2011_09_30_drive_0018 |
| 06 | 2011_09_30_drive_0020 |
| 07 | 2011_09_30_drive_0027 |
| 08 | 2011_09_30_drive_0028 |
| 09 | 2011_09_30_drive_0033 |
| 10 | 2011_09_30_drive_0034 |
the processed KITTI data for training and validation will be placed in folder KITTI_rec_256 as
follows
├── 2011_09_26_drive_0067_sync_02
├── 2011_09_30_drive_0016_sync_02
├── 2011_09_30_drive_0018_sync_02
├── 2011_09_30_drive_0020_sync_02
├── 2011_09_30_drive_0027_sync_02
├── 2011_09_30_drive_0028_sync_02
├── 2011_09_30_drive_0033_sync_02
├── 2011_09_30_drive_0034_sync_02
├── 2011_10_03_drive_0027_sync_02
├── 2011_10_03_drive_0034_sync_02
├── 2011_10_03_drive_0042_sync_02
├── train.txt
└── val.txt
There are images, camera intrinsic parameters, sparse imu values, dense imu values, ground truth poses, sampled imu data in each sub folder. The structure are listed as follows:
├── 000000xxxx.jpg
├── ......
├── cam.txt
├── oxts.csv
├── oxts_ori.csv
├── poses.csv
├── sampled_imu_index1_index2.npy
We list the folder structure of Malaga dataset as follows,
├── 01
├── 02
├── ...
in each folder, there are imu data, images, timestamp arranged as follows,
├── imu
├── left_256
├──── xxxxxxxxxx.xxxxxx.jpg
├──── ......
├──── cam.txt
├── data.csv