- Setup Autoware
- Download any of the KITTI dataset RAW rectified sets and its tracklets (http://www.cvlibs.net/datasets/kitti/raw_data.php). You should have a directory structure that looks like this:
kitti_download_dir/2011_09_26/2011_09_26_drive_0001_sync
├── image_00
│ └── data
├── image_01
│ └── data
├── image_02
│ └── data
├── image_03
│ └── data
├── oxts
│ └── data
└── velodyne_points
└── data
- Download tracklets file for the RAW data.
kitti_download_dir/2011_09_26/2011_09_26_drive_0001_sync
|
~~~~~~~
└──tracklets_labels.xml
-
Make rosbag by using this repo.
-
Set
kitti_data_dirto your kitti data directrory path in launch/benchmark.launch#L104 -
Launch
benchmark.launchin launch directory. -
Play the rosbag made in step 4.
-
After finished recording results, you should get
benchmark_result.txtfile under the kitti data directory.
kitti_download_dir/2011_09_26/2011_09_26_drive_0001_sync
|
~~~~~~~
└──benchmark_result.txt
-
Run
python3 evaluate_tracking.py -
You will see benchmark directory under
kitti data directory.