Loading...
Loading...
Self-driving perception, LiDAR/camera fusion, trajectory prediction, drone perception, robot manipulation
2,023 datasets
Synchronized multimodal driving data was collected in the CARLA simulator using the autopilot feature. The dataset includes RGB images, semantic segmentation, LiDAR point clouds, 2D bounding boxes, and ego-vehicle state data across varied weather and traffic conditions. It was created by immanuelpeter and last updated on 2025-11-29.
Encompassing images for training object detection models to identify drones, primarily rotary-wing UAVs, in various environments. It was curated from 23 open-source datasets and processed through a custom cleaning pipeline. The author is lgrzybowski, and it was last updated in December 2025.
A 2023 drone-based lidar survey of a mixed-conifer forest on Cle Elum Ridge, Washington, conducted on 6 March 2023 by the Natural Hazards Reconnaissance (RAPID) Facility. The dataset provides high-resolution point clouds at approximately 100 points per square meter, detailing vegetation structure and snow-on surface topography in areas with and without fuel-reduction treatments. It complements a 2021 survey to enable pre- and post-treatment comparisons of forest structure and snowpack.
KevinLADLee authored this suite of data collection tools for the CARLA simulator, updated in January 2026, to facilitate the generation of synthetic autonomous driving data. The repository enables researchers to programmatically capture synchronized sensor outputs and vehicle states within simulated urban environments.
AgriLiRa4D is a multi-sensor UAV dataset designed for robust Simultaneous Localization and Mapping (SLAM) in challenging agricultural environments. The dataset was created by Zhihao Zhan, Yuhang Ming, Shaobin Li, and Jie Yuan and was released in 2025. It is hosted on Hugging Face and was last updated on December 6, 2025.
A new dataset for multi-moving-camera tracking, as described in the paper 'Towards Effective Multi-Moving-Camera Tracking: A New Dataset and Lightweight Link Model'. The dataset is hosted on Hugging Face by author ajhcdsgfewgyuyuha and was last updated on December 10, 2025. It is intended to support research into systematic tracking of on-road pedestrians for autonomous vehicle safety.
Facebear's dataset contains approximately 1,500 episodes of robotic cloth folding, collected using an Agilex Aloha robotic arm. It was created for the X-VLA paper and demonstrates a near-perfect success rate in the folding task. The dataset was last updated on Hugging Face in November 2025.
Multispectral driving data categorized into RGB and Short-Wave Infrared (SWIR) bands captured across five adverse weather conditions. The collection includes synchronized image pairs specifically designed to address perception failures in fog, rain, snow, glare, and high-contrast environments.
PHUMA is a physically-grounded humanoid locomotion dataset created by DAVIAN-Robotics. It leverages large-scale human motion data, processed through physics-constrained retargeting to overcome physical artifacts. The dataset was last updated in November 2025.
Waymo Open Dataset provides high-resolution sensor data and 3D labels for autonomous-driving research, released by waymo-research. The collection was last updated in January 2026 and serves as a benchmark for perception and motion prediction tasks.
USC-PSI-Lab released Humanoid Everyday in 2025, providing over 260 tasks across 7 categories for open-world robotic learning. The collection includes between 1,000 and 10,000 video records of humanoid manipulation and locomotion-integrated activities. All data were captured through a human-supervised teleoperation pipeline to ensure high-quality demonstration trajectories.
USC-PSI-Lab released Humanoid Everyday in 2025, providing over 260 tasks across 7 categories for open-world robotic learning. The collection includes between 1,000 and 10,000 video records of humanoid manipulation and locomotion-integrated activities. All data were captured through a human-supervised teleoperation pipeline to ensure high-quality demonstration trajectories.
NVIDIA Corporation created a dataset of 10 synchronized clips of LiDAR and multi-view video, accompanied by corresponding camera and LiDAR poses. The dataset was published on October 23, 2025. It is intended for commercial and non-commercial use under the NVIDIA Autonomous Vehicle Dataset License Agreement.
Ma Lio provides sensor data for asynchronous multiple LiDAR-inertial odometry, developed by minwoo0611 and updated in December 2025. The repository facilitates SLAM research through point-wise inter-LiDAR uncertainty propagation across multiple sensor streams.
Airborne lidar datasets collected in the Arctic Ocean during July 2022 as part of NASA's Calibration and Validation Campaign for ICESat-2. The data, published by researcher Kutalmis Saylam via the Texas Data Repository, were acquired using a Leica Chiroptera 4X system from Thule Air Force Base. It includes two sets of data for near-infrared (1064 nm) and green-wavelength (515 nm) light.
European spatially explicit data on forest canopy fuel load and canopy bulk density at a 1 km² grid resolution. The dataset comprises 4 maps, including the two fuel parameters and their associated uncertainties, generated using a multi-sensor approach integrating GEDI LiDAR, Landsat 8, and PALSAR SAR imagery with machine learning. It was created by Aragoneses, Elena and colleagues, with the supporting publication dated 2025.
Over 10,853 real-world samples and 26,600 synthetic samples for mmWave beam selection in vehicle-to-everything networks. The dataset, created by Muruganandham, Divyadharshini, consists of synchronized LiDAR, camera, and GPS data in diverse scenarios and a high-fidelity digital twin. It was last updated on 2025-10-15.
SynLiDAR is a synthetic LiDAR sequential point cloud dataset featuring point-wise semantic annotations, developed by xiaoaoran and published at AAAI 2022. It provides large-scale simulated sensor data specifically designed for 3D scene understanding and domain adaptation research.
Volume 1 of the Texas Coastal Hazards Atlas contains geospatial data for hurricane preparedness and coastal management along the state's southeast coast. The dataset, authored by James Gibeaut and hosted by the Texas Data Repository, includes sections on bay erosion, hurricane surge modeling, shoreline projections, and environmental sensitivity. It was last updated on October 15, 2025.
A single oral history interview conducted by the Veteranen Instituut, focusing on a veteran's experiences after World War II. The interview covers his artillery service, deployment to the Dutch Indies aboard the Johan de Witt, arrival in Batavia, and subsequent military actions including the First Police Action and the Darul Islam movement. The dataset was last updated on October 24, 2025, and is part of the DANS Data Station Social Sciences and Humanities Collection.