Loading...
Loading...
Self-driving perception, LiDAR/camera fusion, trajectory prediction, drone perception, robot manipulation
2,016 datasets
Two experimental missions from April 2015 to September 2016 collected near-surface environmental data in the Bering Sea using autonomous Saildrone vehicles. The dataset includes over 20 variables such as water temperature, salinity, chlorophyll-a, ocean currents, wind speed, air temperature, and pressure, recorded by instruments like thermosalinographs, ADCPs, and fluorometers. These experimental data were collected by NOAA's Pacific Marine Environmental Laboratory and Saildrone LLC under a CRADA to assess the platform's viability.
CUAVerifierBench is an evaluation benchmark for verifiers of computer-using agents, created by Microsoft. It contains human-annotated trajectories of agent interactions to judge task completion. The dataset was last updated on 2026-04-21.
NOAA's collection of lidar projects from multiple sources, geographically focused on the coastal areas of the United States. The data consists of 3D point clouds, with each point attributed by object type, reflection intensity, and system metadata. Data is provided in both Entwine Point Tiles (EPT) and LAZ formats, using coordinate reference systems appropriate for each dataset's location.
EgoTraj-Bench is the first real-world benchmark for pedestrian trajectory prediction under ego-centric noisy observations. It pairs noisy first-person-view derived trajectories with clean bird's-eye-view ground truth, built upon the TBD dataset. The dataset was uploaded by ZoeyLIU1999 and last updated on April 22, 2026.
A collection of 1,104 unique runs of multi-sensor data collected at a controlled roadway intersection for training intersection safety algorithms. Data was gathered from 20 roadside sensors, including visual and thermal cameras, LiDAR, and radar, at a test facility in McLean, VA from October 2023 through March 2024. The full dataset is approximately 1 TB in size.
Featuring 41 unique runs of multi-sensor data from a controlled intersection test facility, totaling approximately 30 GB. It includes labeled validation data with 3D bounding boxes and conflict/no-conflict labels for road users, collected from 20 roadside sensors including cameras, LiDAR, and radar. The data was gathered by the U.S. Department of Transportation at the TFHRC Smart Intersection in McLean, VA from October 2023 through March 2024.
A Sim2Real Transfer Framework for Vision-Based ADAS Using Object Detection. The dataset likely contains synthetic and real-world imagery for training and validating perception models. Its specific scale, creation date, and authorship are not detailed in the available metadata.
Raw 4K drone footage captures aquatic environments. The dataset is hosted on Kaggle, but details on the number of videos, collection dates, and the author are unknown. The content consists of unprocessed aerial video recordings.
Aerial drone imagery extracted at 1 frame per second for urban planning applications. The dataset is hosted on Kaggle and its columns suggest it likely contains image data. Metadata is minimal; actual content requires verification after download.
Slamet Riyadi Elementary School data published on Kaggle. The dataset likely contains information related to primary education in Bandung, Indonesia. Specific details such as row count, columns, and file format are not provided in the metadata.
NOAA OCM's 2022 dataset provides a 1-meter resolution topobathy Digital Elevation Model for approximately 219 square miles along the shores of Green Bay, Wisconsin. The data were collected by Dewberry using a CZMIL Super Nova system from July 19 to August 16, 2022. It consists of 2,630 individual 500m x 500m GeoTiff files.
A lidar point cloud dataset covering approximately 85 square miles along the shores of Tarpon Springs, Florida. The data was collected by NOAA using a Riegl VQ880G sensor from February 11 to 14, 2016. It includes 1,224 tiles of topobathy data in LAS 1.2 format with classifications for ground, water surface, and bathymetric bottom.
A lidar point cloud dataset covering approximately 35 square miles along the shores of Dog Island, Florida. The data were collected by the National Oceanic and Atmospheric Administration using a Riegl VQ880G sensor from April 14 to 16, 2017. It includes 616 tiles of 500 m x 500 m data classified for topographic and bathymetric features.
616 lidar tiles covering approximately 35 square miles along the shores of Dog Island, Florida. The data were collected by the National Oceanic and Atmospheric Administration using a Riegl VQ880G sensor from April 14 to 16, 2017. The point cloud data is in LAS 1.2 format and includes classifications for ground, water surface, bathymetric bottom, and water column.
From October 18 to December 3, 2022, Dewberry collected topobathy lidar data using a CZMIL Super Nova system for the National Oceanic and Atmospheric Administration. The dataset covers approximately 1,373 square miles along Florida's Big Bend coast and contains 17,639 classified 500-meter tiles. The South Block subset comprises 9,585 of these tiles.
Coastal Maine topobathymetric lidar data collected by NV5 for NOAA between October 2022 and December 2023. The dataset includes 120 missions flown with Leica Chiroptera Hawkeye 4X systems, providing point cloud data in LAS format with intensity, returns, and scan angle. NOAA reviewed the data, leading to corrections and normalized green laser intensity values for depth, with a full redelivery of LAS files and normalized intensity rasters.
Florida's Big Bend coastal region from Bayonet Point to Cedar Key was surveyed using a CZMIL Super Nova system between October 18 and December 3, 2022. The dataset contains approximately 1,373 square miles of topobathy data, delivered as 118 5,000 m x 5,000 m DEM tiles. It was collected by Dewberry for the National Oceanic and Atmospheric Administration.
Big Bend Wildlife Management Area, Florida, coastal data covering approximately 1,373 square miles. The National Oceanic and Atmospheric Administration collected these topobathy lidar point clouds using a CZMIL Super Nova system from October 27 to November 29, 2022. The dataset includes 17,482 tiles, with the North Block subset containing 8,054 individual 500-meter by 500-meter tiles.
Virginia's Roanoke, Rappahannock, Shenandoah, Appomattox, and James Rivers are covered by approximately 195 square miles of topobathy lidar data. The National Oceanic and Atmospheric Administration collected these data using a CZMIL Super Nova system from October 2023 through November 2024. The dataset consists of 272 digital elevation model tiles in LAS 1.4 format, classified for features like ground, water surface, and bathymetric bottom.
Contour lines at a 1-meter vertical interval cover the entire Ipswich City Council authority area. The data was compiled by the Ipswich City Council from a 2019 Airborne Laser Scanning (LiDAR) project. It provides a detailed topographic model of the region.