Publicly available Lidar datasets for Autonomous Vehicle use-cases
Lidar datasets you can start using right away.
| Dataset | License | Links | Description |
|---|---|---|---|
KITTI |
CC-BY-NC-SA-3.0 | Raw data3D cuboid annotationsPanoptic segmentation of images |
GeographyKarlsruhe, Germany
Sensors2 grayscale cameras, 2 color cameras, Velodyne LiDAR, GPS/IMU
Data formatBIN, PNG and JSON files
Annotations3D/2D object detection, tracking, semantic segmentation
DetailsA pioneering dataset widely used for benchmarking autonomous driving algorithms and ML models. It contains 7481 point cloud files with ~80K labeled objects. You need to create an account and wait for approval before you can access the dataset. Note that the 3D cuboid annotations are only labeled top-down, meaning they are not accurate in Z-axis. Moreover, only objects visible in the cameras are labeled.
|
Nuscenes |
CC BY-NC-SA 4.0 | Download page for main datasetNuImagesDevkit |
GeographyBoston (USA), Singapore
Sensors6 cameras, 5 RADARs, 1 LiDAR, GPS/IMU
Data formatBIN, PNG and JSON files
Annotations3D bounding boxes, object tracking, semantic maps
DetailsNuscenes is a high-resolution, multi-modal sensor data collected in diverse weather and lighting conditions. It covers a total of 35K labeled keyframes across 850 scenes. You need to create an account before you can access the dataset.
|
ONCE |
CC BY-NC-SA 4.0 | Download pageData formatDevkit |
GeographyChina (urban areas)
SensorsLiDAR and 7 RGB cameras
Data formatPCD, PNG and JSON files
AnnotationsPartially annotated for 3D object detection
DetailsONCE is one of the largest autonomous driving datasets, focusing on real-world urban scenes. It covers 1 Million LiDAR frames and 7 Million camera images, with 15K fully annotated scenes.
|
WAYMO |
Custom Non-commercial license | Download pageResearch paperDevkit |
GeographyUSA (California, Arizona)
Sensors5 LiDARs and 5 cameras
Data formatSharded TFRecord format files containing protocol buffer data
AnnotationsWide variety covering tracked 2D/3D objects, key-points & segmentation for 2D/3D
DetailsThe Waymo Open Dataset covers a wide variety of annotations over a large number of frames. It is composed of three datasets - the Perception Dataset with high resolution sensor data and labels for 2,030 segments. The Motion Dataset with object trajectories and corresponding 3D maps for 103,354 segments, and the End-to-End Driving Dataset with camera images providing 360-degree coverage and routing instructions for 5,000 segments.
|
Wayve Scenes |
Custom Non-commercial license | Download pageResearch paperDevkit |
GeographyUS & UK (urban and suburban areas)
Sensors5 high resolution cameras and GPS
Data formatRaw and labeled masks in PNG format
AnnotationsMostly unlabeled - some segmentation and trajectory data is available
DetailsThe WayveScenes dataset targets applications of novel view synthesis. It contains 101 diverse scenes covering 101K images. Note that most of these images are unlabeled
|
Argoverse 2 |
CC BY-NC-SA 4.0 | Download pageResearch paperDevkit |
Geography6 USA cities
Sensors2 Lidar sensors, 7 ring cameras and 2 stereo cameras
Data formatPoint cloud files and annotations as as Apache feather files, Raw images as PNG files
AnnotationsAnnotated HD maps, ground points, tracked 3D cuboids, motion forecasting and map change detection data
DetailsArgoverse 2 is a comprehensive dataset for a number of tasks including 3D object detection, tracking, detecting changes in HD maps and motion forecasting. The sensor dataset contains 1000 labeled scenarios of 15 seconds each.
|
LIVOX |
GNU General Public License v3.0 (GPLv3) | Download pageDevkit |
GeographyChina (urban environments)
Sensors5 Horizon lidars and 1 Tele-15 lidar
Data formatPoint cloud files as custom BIN and annotations as JSON files
Annotations3D bounding box and semantic segmentation annotations
DetailsLIVOX is a simulated dataset containing 14K frames aimed at helping ressearchers and LIVOX customers perform tasks such as 3D object detection and semantic segmentation.
|
A2D2 |
CC BY-ND 4.0 | Download pageResearch paperDevkit |
GeographyGermany (3 cities)
Sensors5 lidars and 6 cameras
Data formatPoint cloud files as npz(numpy zip) files, images in PNG format and annotations as JSON files
AnnotationsLidar and Image semantic segmentation and 3D cuboid annotations
DetailsA2D2 (Audi Autonomous Driving Dataset) contains 40K images and point cloud files labeled for semantic segmentation, while 12K point cloud frames labeled for 3D object detection.
|
ApolloScape |
Custom non-commercial license | Download pageDevkit |
GeographyBeijing, China
Sensors2 lidars
Data formatPoint cloud files as PCD files, images in PNG format and annotations as JSON files
Annotations3D cuboid annotations for Lidar, disparity maps for sterio and lane detection
DetailsApolloscape consists of datasets for 3D object detection, lane detection, disparity maps for sterio iamges, scene inpanting anf trajectory prediction.
|
WADS |
CC BY 4.0 | Download pageResearch paperData format details |
GeographyMichigan, US
Sensors1 lidar
Data formatPoint cloud files as BIN files, annotations in KITTI PLY format
AnnotationsPantopic segmentation of Lidar data
DetailsWADS (Winter Adverse Driving dataSet) consists of around 2100 point cloud scans fully labeled for pantoptic segmentation. It captures lidar data in falling snow and accumulated snow along with all the semanticKITTI classes.
|
Toronto-3D |
CC BY-NC 4.0 | Download pageResearch paperDevkit |
GeographyToronto, Canada
Sensors1 lidar
Data formatPoint cloud files as BIN files, annotations in PLY format
AnnotationsSemantic segmentation of Lidar data
DetailsToronto-3D is a large-scale urban outdoor point cloud dataset covering approximately 1 km of road and consisting of about 78.3 million labeled points.
|
Want to develop your own custom dataset?
Whether you are looking for Multi-sensor Annotation tool or a complete annotation solution, we can help! Please tell us what you are looking for and we will get back to you within 24 hours.
