The Newer College Dataset: Handheld LiDAR, Inertial and Vision with Ground Truth

Abstract

In this paper we present a large dataset with a variety of mobile mapping sensors collected using a handheld device carried at typical walking speeds for nearly 2.2 km through New College, Oxford. The dataset includes data from two commercially available devices -a stereoscopic-inertial camera and a multi-beam 3D LiDAR, which also provides inertial measurements. Additionally, we used a tripod-mounted survey grade LiDAR scanner to capture a detailed millimeteraccurate 3D map of the test location (containing ∼290 million points). Using the map we inferred centimeter-accurate 6 Degree of Freedom (DoF) ground truth for the position of the device for each LiDAR scan to enable better evaluation of LiDAR and vision localisation, mapping and reconstruction systems. This ground truth is the particular novel contribution of this dataset and we believe that it will enable systematic evaluation which many similar datasets have lacked. The dataset combines both built environments, open spaces and vegetated areas so as to test localization and mapping systems such as vision-based navigation, visual and LiDAR SLAM, 3D LIDAR reconstruction and appearance-based place recognition. The dataset is available at: ori.ox.ac.uk/datasets/newer-college-dataset

Binary Graph Descriptor for Robust Relocalization on Heterogeneous Data

“…To demonstrate the capacity of our BiG descriptor, we relied on multiple public datasets : i) Synthia Dataset [33], a synthetic dataset across different environmental conditions; ii) Newer College Dataset [29], a real recorded dataset providing infrared images and Lidar data; iii) RobotCar Seasons Dataset [34,35], a dataset across different weather and time conditions by a car-mounted camera.…”

“…(left lower) ; k-th order neighbour regions are highlighted in the image (right lower). Image from [29].…”

VDBFusion: Flexible and Efficient TSDF Integration of Range Sensor Data

“…In addition to these assumptions, SuperEight runs at three frames per second (fps) and is thus slower than the LiDAR sensor frame rate. In contrast, our system can efficiently fuse scans at 30 fps under the same constraints and using the same dataset [ 45 ], being, therefore, 10 times more efficient. Lastly, there is currently no information on how this extension [ 1 , 34 ] performs on datasets other than the one used in the original publication [ 45 ].…”

“…In contrast, our system can efficiently fuse scans at 30 fps under the same constraints and using the same dataset [ 45 ], being, therefore, 10 times more efficient. Lastly, there is currently no information on how this extension [ 1 , 34 ] performs on datasets other than the one used in the original publication [ 45 ]. For generating high-fidelity maps, Vizzo et al [ 40 ] employ Poisson surface reconstruction [ 46 ].…”

“… Qualitative results on the Newer College Dataset [ 45 ] with given parameters, dataset statistics, and size of the dataset in respect to spatial extent and memory. …”