Dataset for "Parameter Reduction and Optimisation for Point Cloud and Occupancy Mapping Algorithms"
In the applications of occupancy mapping, the scene to be explored is normally large and objects are of irregular shapes. In this case, it is difficult to evaluate the performance of a mapping approach since the ground truth can hardly be obtained. This dataset aims to provide small measured scenes with ground truths for evaluation purposes.
Videos are recorded in .svo files. These files can be opened with the tools provided in ZED SDK which can be downloaded at Stereo Labs website (https://www.stereolabs.com). Images in the videos can be extracted using the API in ZED SDK and the camera parameters can also be accessed by ZED API.
Ground truth files are in .ot format. These files can be viewed by octovis package. To install octovis, run 'sudo apt-get install octovis' in Ubuntu.
Camera trajectories produced by ORB-SLAM and keyframe poses produced by ORB-SLAM are in .txt files. Most text editors can open files of this type.
Cite this dataset as:
Miao, Y.,
Hunter, A.,
Georgilas, I.,
2021.
Dataset for "Parameter Reduction and Optimisation for Point Cloud and Occupancy Mapping Algorithms".
Bath: University of Bath Research Data Archive.
Available from: https://doi.org/10.15125/BATH-00594.
Export
Data
VoronoiBuildingBoxI.svo
application/octet-stream (3GB)
Creative Commons: Attribution 4.0
VoronoiBuildingBoxO.svo
application/octet-stream (3GB)
Creative Commons: Attribution 4.0
VoronoiBuildingBoxT.svo
application/octet-stream (3GB)
Creative Commons: Attribution 4.0
VoronoiBuildingBoxL.svo
application/octet-stream (3GB)
Creative Commons: Attribution 4.0
VoronoiBuildingBoxS.svo
application/octet-stream (2GB)
Creative Commons: Attribution 4.0
PlainBuildingBoxI.svo
application/octet-stream (2GB)
Creative Commons: Attribution 4.0
PlainBuildingBoxO.svo
application/octet-stream (2GB)
Creative Commons: Attribution 4.0
PlainBuildingBoxT.svo
application/octet-stream (2GB)
Creative Commons: Attribution 4.0
PlainBuildingBoxL.svo
application/octet-stream (2GB)
Creative Commons: Attribution 4.0
PlainBuildingBoxS.svo
application/octet-stream (2GB)
Creative Commons: Attribution 4.0
VoronoiParkingLotBoxI.svo
application/octet-stream (2GB)
Creative Commons: Attribution 4.0
VoronoiParkingLotBoxO.svo
application/octet-stream (2GB)
Creative Commons: Attribution 4.0
VoronoiParkingLotBoxT.svo
application/octet-stream (2GB)
Creative Commons: Attribution 4.0
VoronoiParkingLotBoxL.svo
application/octet-stream (2GB)
Creative Commons: Attribution 4.0
VoronoiParkingLotBoxS.svo
application/octet-stream (2GB)
Creative Commons: Attribution 4.0
PlainParkingLotBoxI.svo
application/octet-stream (1GB)
Creative Commons: Attribution 4.0
PlainParkingLotBoxO.svo
application/octet-stream (2GB)
Creative Commons: Attribution 4.0
PlainParkingLotBoxT.svo
application/octet-stream (1GB)
Creative Commons: Attribution 4.0
PlainParkingLotBoxL.svo
application/octet-stream (1GB)
Creative Commons: Attribution 4.0
PlainParkingLotBoxS.svo
application/octet-stream (1GB)
Creative Commons: Attribution 4.0
GroundTruth.zip
application/zip (421kB)
Creative Commons: Attribution 4.0
CameraTrajectories.zip
application/zip (1MB)
Creative Commons: Attribution 4.0
KeyframePoses.zip
application/zip (95kB)
Creative Commons: Attribution 4.0
Creators
Yu Miao
University of Bath
Alan Hunter
University of Bath
Ioannis Georgilas
University of Bath
Contributors
University of Bath
Rights Holder
Coverage
Geographical coverage:
The Parade and East Extension Car Park, University of Bath
Documentation
Data collection method:
This dataset contains 20 video files used for occupancy mapping. The dataset was recorded with a ZED stereo camera (Stereo Labs, USA) in two environments at University of Bath, in front of buildings and in a parking lot. Targets in the dataset are two boxes whose external surfaces are either with a plain brown cardboard texture or are covered with randomly generated Voronoi diagrams. The average size of the polygons in each diagram is about 3 cm x 3 cm. The colour of each polygon is randomly generated. The Voronoi diagrams are printed in 300 DPI on the posters of A0 size and cropped to cover the surfaces of boxes. Two boxes are arranged in a layout as one of five free tetrominoes, i.e., I, O, T, L and S, in the Tetris game. The camera obits the boxes twice in a anti-clock direction to capture videos. Ground truths are also included in this dataset package.
Technical details and requirements:
The device for data collection is a ZED camera from Stereo Labs, USA. Software is ZED SDK which is available at https://www.stereolabs.com.
Funders
University of Bath
https://doi.org/10.13039/501100000835
Publication details
Publication date: 14 June 2021
by: University of Bath
Version: 1
DOI: https://doi.org/10.15125/BATH-00594
URL for this record: https://researchdata.bath.ac.uk/id/eprint/594
Related papers and books
Miao, Y., Hunter, A., and Georgilas, I., 2021. Parameter Reduction and Optimisation for Point Cloud and Occupancy Mapping Algorithms. Sensors, 21(21), 7004. Available from: https://doi.org/10.3390/s21217004.
Miao, Y., Hunter, A., and Georgilas, I., 2021. An Occupancy Mapping Method Based on K-Nearest Neighbours. Sensors, 22(1), 139. Available from: https://doi.org/10.3390/s22010139.
Contact information
Please contact the Research Data Service in the first instance for all matters concerning this item.
Contact person: Yu Miao
Faculty of Engineering & Design
Mechanical Engineering
