The files mentioned below can also be downloaded via OpenDataLab. It is recommended to use provided command line interface (CLI) for acceleration.
| Subset | Split | Google Drive  |
Baidu Yun |
md5 | Size |
|---|---|---|---|---|---|
| sample | OpenLane-V2 | sample | sample | 21c607fa5a1930275b7f1409b25042a0 | ~300M |
| subset_A | OpenLane-V2 | info | info | 95bf28ccf22583d20434d75800be065d | ~8.8G |
| Map Element Bucket | info | info | 1c1f9d49ecd47d6bc5bf093f38fb68c9 | ~240M | |
| Image (train) | image_0 | image_0 | 8ade7daeec1b64f8ab91a50c81d812f6 | ~14.0G | |
| image_1 | image_1 | c78e776f79e2394d2d5d95b7b5985e0f | ~14.3G | ||
| image_2 | image_2 | 4bf09079144aa54cb4dcd5ff6e00cf79 | ~14.2G | ||
| image_3 | image_3 | fd9e64345445975f462213b209632aee | ~14.4G | ||
| image_4 | image_4 | ae07e48c88ea2c3f6afbdf5ff71e9821 | ~14.5G | ||
| image_5 | image_5 | df62c1f6e6b3fb2a2a0868c78ab19c92 | ~14.2G | ||
| image_6 | image_6 | 7bff1ce30329235f8e0f25f6f6653b8f | ~14.4G | ||
| Image (val) | image_7 | image_7 | c73af4a7aef2692b96e4e00795120504 | ~21.0G | |
| Image (test) | image_8 | image_8 | fb2f61e7309e0b48e2697e085a66a259 | ~21.2G | |
| SD Map | sdmap | sdmap | de22c7be880b667f1b3373ff665aac2e | ~7M | |
| subset_B | OpenLane-V2 | info | info | 27696b1ed1d99b1f70fdb68f439dc87d | ~7.7G |
| Image (train) | image_0 | image_0 | 0876c6b2381bacedeb3be16e57c7d59b | ~3.4G | |
| image_1 | image_1 | ecdec8ff8c72525af322032a312aad10 | ~3.3G | ||
| image_2 | image_2 | b720bf7fdf0ebd44b71beffc84722359 | ~3.3G | ||
| image_3 | image_3 | ac3bc9400ade6c47c396af4b12bbd0e0 | ~3.4G | ||
| image_4 | image_4 | fa4c4a04b5ad3eac817e6368047d0d89 | ~3.5G | ||
| image_5 | image_5 | 19d2cc92514e65270779e405d3a93c61 | ~3.6G | ||
| image_6 | image_6 | d4f56c562f11a6bcc918f2d20441c42c | ~3.3G | ||
| Image (val) | image_7 | image_7 | 443045d7a3faf5998af27e2302d3503e | ~5.0G | |
| Image (test) | image_8 | image_8 | 6ecb7a9e866e29ed73d335c2d897f50e | ~5.4G |
OpenLane-V2contains annotations for the initial task of OpenLane Topology.Map Element Bucketcontains annotations for the task of Driving Scene Topology.ImageandSD Mapserves as sensor inputs.
For files in Google Drive, you can use the following command by replacing [FILE_ID] and [FILE_NAME] accordingly:
wget --load-cookies /tmp/cookies.txt "https://docs.google.com/uc?export=download&confirm=$(wget --quiet --save-cookies /tmp/cookies.txt --keep-session-cookies --no-check-certificate 'https://docs.google.com/uc?export=download&id=[FILE_ID]' -O- | sed -rn 's/.*confirm=([0-9A-Za-z_]+).*/\1\n/p')&id=[FILE_ID]" -O [FILE_NAME]The dataset is preprocessed into pickle files representing different collections, which then be used for training models or evaluation:
cd data
python OpenLane-V2/preprocess.py For using the SD Map and Map Element Bucket, please specify the with_sd_map parameter and run python OpenLane-V2/preprocess-ls.py respectively.
The hierarchy of folder OpenLane-V2/ is described below:
└── OpenLane-V2
├── train
| ├── [segment_id]
| | ├── image (Image)
| | | ├── [camera]
| | | | ├── [timestamp].jpg
| | | | └── ...
| | | └── ...
| | ├── sdmap.json (SD Map)
| | └── info
| | ├── [timestamp].json (OpenLane-V2)
| | ├── [timestamp]-ls.json (Map Element Bucket)
| | └── ...
| └── ...
├── val
| └── ...
├── test
| └── ...
├── data_dict_example.json
├── data_dict_subset_A.json
├── data_dict_subset_B.json
├── openlanev2.md5
├── preprocess.py
└── preprocess-ls.py
[segment_id]specifies a sequence of frames, and[timestamp]specifies a single frame in a sequence.image/contains images captured by various cameras, andinfo/contains meta data and annotations of a single frame.data_dict_[xxx].jsonnotes the split of train / val / test under the subset of data.
The sdmap.json comprises three types of SD map elements that can be used as sensor inputs.
[
{
'points': <list> -- list of 2D points in the BEV space
'category': <str> -- type of the SD map element
'road',
'cross_walk',
'side_walk',
},
...
]
The json files under the info/ folder contain meta data and annotations for each frame.
Each file is formatted as follows:
{
'version': <str> -- version
'segment_id': <str> -- segment_id
'meta_data': {
'source': <str> -- name of the original dataset
'source_id': <str> -- original identifier of the segment
}
'timestamp': <int> -- timestamp of the frame
'sensor': {
[camera]: { <str> -- name of the camera
'image_path': <str> -- image path
'extrinsic': <dict> -- extrinsic parameters of the camera
'intrinsic': <dict> -- intrinsic parameters of the camera
},
...
}
'pose': <dict> -- ego pose
'annotation': <dict> -- anntations for the current frame
}
For a single frame, annotations are formatted as follow:
{
'lane_centerline': [ (n lane centerlines in the current frame)
{
'id': <int> -- unique ID in the current frame
'points': <float> [n, 3] -- 3D coordiate
'confidence': <float> -- confidence, only for prediction
},
...
],
'traffic_element': [ (k traffic elements in the current frame)
{
'id': <int> -- unique ID in the current frame
'category': <int> -- traffic element category
1: 'traffic_light',
2: 'road_sign',
'attribute': <int> -- attribute of traffic element
0: 'unknown',
1: 'red',
2: 'green',
3: 'yellow',
4: 'go_straight',
5: 'turn_left',
6: 'turn_right',
7: 'no_left_turn',
8: 'no_right_turn',
9: 'u_turn',
10: 'no_u_turn',
11: 'slight_left',
12: 'slight_right',
'points': <float> [2, 2] -- top-left and bottom-right corners of the 2D bounding box
'confidence': <float> -- confidence, only for prediction
},
...
],
'topology_lclc': <float> [n, n] -- adjacent matrix among lane centerlines
'topology_lcte': <float> [n, k] -- adjacent matrix between lane centerlines and traffic elements
}
idis the identifier of a lane centerline or traffic element and is consistent in a sequence. For predictions, it can be randomly assigned but unique in a single frame.topology_lclcandtopology_lcteare adjacent matrices, where row and column are sorted according to the order of the listslane_centerlineandtraffic_element. It is a MUST to keep the ordering the same for correct evaluation. For ground truth, only 0 or 1 is a valid boolean value for an element in the matrix. For predictions, the value varies from 0 to 1, representing the confidence of the predicted relationship.#lane_centerlineand#traffic_elementare not required to be equal between ground truth and predictions. In the process of evaluation, a matching of ground truth and predictions is determined.
In the Map Element Bucket, we reformulate the annotation files to include additional labels.
{
'lane_segment': [ (i lane segments in the current frame)
{
'id': <int> -- unique ID in the current frame
'centerline': <float> [n, 3] -- 3D coordiate
'left_laneline': <float> [n, 3] -- 3D coordiate
'left_laneline_type': <int> -- type of the left laneline
0: 'none',
1: 'solid',
2: 'dash',
'right_laneline': <float> [n, 3] -- 3D coordiate
'right_laneline_type': <int> -- type of the right laneline
'is_intersection_or_connector': <bool> -- whether the lane segment is in a intersection or connector
'confidence': <float> -- confidence, only for prediction
},
...
],
'traffic_element': [ (j traffic elements in the current frame)
{
'id': <int> -- unique ID in the current frame
'category': <int> -- traffic element category
1: 'traffic_light',
2: 'road_sign',
'attribute': <int> -- attribute of traffic element
0: 'unknown',
1: 'red',
2: 'green',
3: 'yellow',
4: 'go_straight',
5: 'turn_left',
6: 'turn_right',
7: 'no_left_turn',
8: 'no_right_turn',
9: 'u_turn',
10: 'no_u_turn',
11: 'slight_left',
12: 'slight_right',
'points': <float> [2, 2] -- top-left and bottom-right corners of the 2D bounding box
'confidence': <float> -- confidence, only for prediction
},
...
],
'area': [ (k areas in the current frame)
{
'id': <int> -- unique ID in the current frame
'category': <int> -- area category
1: 'pedestrian_crossing',
2: 'road_boundary',
'points': <float> [n, 3] -- 3D coordiate
'confidence': <float> -- confidence, only for prediction
},
...
],
'topology_lsls': <float> [n, n] -- adjacent matrix among lane segments
'topology_lste': <float> [n, k] -- adjacent matrix between lane segments and traffic elements
}