File(s) not publicly available
Reason: This is a large dataset that cannot be downloaded directly from the record page. Please contact opendata@cardiff.ac.uk to request access.
MirrEnv: a benchmarking dataset for visual SLAM in mirror environments
The Mirror Environments (MirrEnv) dataset is made up of RGBD image sequences with ground-truth camera localization data. Three sizes of mirrors were used (small, medium, large), where some sequences had the mirror visible, covered by a green card, or had the mirror removed from the scene entirely (7 variations in total). Each of these 7 mirror presence varieties was combined with 7 pre-calculated robot arm trajectories, giving a total of 49 sequences in total. Binary masks for the mirror region are provided for every 10th frame of sequences containing mirrors.
The dataset structure is as below. Each sequence is labeled as Trj_X1_X2_X3_X4, where X1 is the numerical index across all 49 sequences; X2 is the trajectory name; X3 indicates the mirror size; X4 indicates if the mirror is visible (W) or covered (C). If there is no mirror, then X3 and X4 are replaced with No_Mirror.
Trj_X1_X2_X3_X4
|----calib
| |----images
| | |----depth
| | | |----#timestamp#.png (16-bit greyscale image, 1280x720)
| | | |----......
| | |----rgb
| | | |----#timestamp#.png (24-bit rgb image, 1280x720)
| | | |----......
| | |----rs_intrinsics.xml
| |----poses
| | |----#timestamp#.txt (4x4 rigid body transformation relative to robot base)
| | |----......
| |----calib_X.txt
| |----DepthFactor.txt
|
|----trj
| |----images
| | |----depth
| | | |----#timestamp#.png (16-bit greyscale image, 640x480)
| | | |----......
| | |----rgb
| | | |----#timestamp#.png (24-bit rgb image, 640x480)
| | | |----......
| | |----masks (only available in sequences containing mirrors, ie. X4 = W)
| | | |----#timestamp#.png (8-bit binary image, 640x480)
| | | |----......
| | |----c_names.txt (filenames of rgb frames - timestamp in seconds since Unix epoch)
| | |----d_names.txt (filesnames of depth frames - timestamp in seconds since Unix epoch)
| | |----depth.mp4 (original video of depth frames)
| | |----rgb.avi (original video of rgb frames)
| |----poses
| | |----Qposes.txt (Camera pose relative to the base of the robot base, XYZ position in metres and unit quaternion pose)
The RGB frames and depth frames have slightly differing timestamps. During experiments, they were associated together using the associate.py Python script available from the TUM RGBD dataset tools.
Research results based upon these data are published at https://doi.org/10.1007/s41095-022-0329-x
Funding
DTP 2020-2021 Cardiff University (2020-10-01 - 2025-09-30); Graham, Kim. Funder: Engineering and Physical Sciences Research Council
Reflection Aware Visual Simultaneous Localization and Mapping (RA-vSLAM) (2020-10-01 - 2024-09-30); Herbert, Peter. Funder: Engineering and Physical Sciences Research Council
History
Specialist software required to view data files
Images can be opened in most image viewers, or even just Python; .txt and .xml can be readable with a text editor, and the .mp4 and .avi can be opened in any media player with a suitable codec installed.Usage metrics
