V4RL Datasets
At V4RL we recularly release our datasets, code, and the setups that we used to encourage benchmarking. Here you can find all of our datasets, while this link details all our code releases, and this link lists our released setups.
Aerial Single-view Depth Completion
In this work, we propose a powerful methodology for depth completion and uncertainty estimation approach that better handles the challenges of mapping from aerial vehicles, such as large viewpoint and depth variations, and limited computing resources. We publicly release our code, datasets and simulator to aid benchmarking. more
Wide-baseline Place Recognition Dataset
This dataset provides sequences from two synthetic and two real outdoor scenes with visual, inertial, and ground-truth information, specifically recorded for wide-baseline place recognition tasks more...
Urban Place Recognition Dataset
This dataset provides visual and inertial information as well as manually-annotated ground-truth for place recognition tasks, outdoors in urban scenes, dataset posing significant challenges in viewpoint, illumination and situational changes.This dataset is freely available more ...
Aerial Inspection Dataset
This dataset contains visual and inertial sequences recorded from the ground and the air (using a small rotorcraft) while moving around a building. This data is captured with a hardware-synchronised sensor and ground-truth of the scene has been captured using a laster scanner. This dataset is freely available more ...
Dense Reconstruction Dataset
A pipe structure and a desk sequence provide visual, inertial & depth data, as well as ground truth for both the pose of the sensor head and the 3D scene. This dataset is publically available. more ...