Awesome visual place recognition (VPR) datasets

This repository provides a curated list of awesome datasets for Visual Place Recognition (VPR), which is also called loop closure detection (LCD). It is a significant component in V-SLAM (Visual Simultaneous Localization and Mapping) systems.

We will update the list if new VPR dataset is presented.

Our survey paper on deep learning-based visual place recognition contains detailed information about these datasets:

Xiwu Zhang, Lei Wang, and Yan Su. Visual Place Recognition: A Survey From Deep Learning Perspective. Pattern Recognition, November 2020, doi: https://doi.org/10.1016/j.patcog.2020.107760.

Datasets used in VPR.

This list summarizes public available datasets that can be used for visual place recognition. We devide them into multiple categories according to different research topics, for example, long-term, across-seasons, semantic VPR and so on.

Topic	Name	Year	Image type	Environment	Illumination	Viewpoint	Ground Truth	Labels	Extra Information
Generic	New College and City Centre [1]	2008	RGB	Outdoor	slight	✔️	✔️	✔️	GPS
	New College Vision and Laser [2]	2009	Gray.	Outdoor	slight	✔️	✔️		GPS, IMU, LiDAR
	Rawseeds	2006	RGB	Indoor/Outdoor		✔️	✔️		GPS, LiDAR
	Ford Campus	2011	RGB	Urban	slight		✔️		GPS, IMU, LiDAR
	Malaga Parking 6L	2009	RGB	Outdoor			✔️		GPS, IMU, LiDAR
	KITTI Odometry	2012	Gray./ RGB	Urban	slight		✔️		GPS, IMU, LiDAR
Long-term	St. Lucia	2010	RGB	Urban	✔️	slight			GPS
	COLD	2009	RGB	Indoor	✔️	✔️	✔️	✔️	LiDAR
	Oxford RobotCar	2017	RGB	Urban	✔️		✔️		GPS, IMU, LiDAR
	Gardens Point Walking	2014	RGB	Indoor/ Outdoor	✔️	✔️			-
	MSLS	2020	RGB	Urban	✔️	✔️	✔️		GPS
Across seasons	Nurburgring and Alderley	2012	RGB	Urban	✔️		✔️	✔️	-
	Nordland	2013	RGB	Outdoor	✔️		✔️		GPS
	CMU	2011	RGB	Urban	✔️	✔️	✔️		GPS
	Freiburg (FAS)	2014	RGB	Urban	✔️		✔️	✔️	GPS
	VPRiCE	2015	RGB	Outdoor	✔️	✔️			-
RGB-D	TUM RGB-D	2012	RGB-D	Indoor		✔️	✔️		IMU
	Microsoft 7-Scenes	2013	RGB-D	Indoor		✔️	✔️	✔️	-
	ICL-NUIM	2014	RGB-D	Indoor		✔️	✔️		-
Semantic	KITTI Semantic	2019	RGB	Urban			✔️	✔️	GPS, IMU, LiDAR
	Cityscapes	2016	RGB	Urban			✔️	✔️	GPS
	CSC	2019	RGB	Outdoor	✔️		✔️		LiDAR
Train networks	Cambridge Landmarks	2015	RGB	Outdoor	✔️	✔️	✔️	✔️	-
	Pittsburgh250k	2013	RGB	Urban	✔️	✔️	✔️	✔️	GPS
	Tokyo 24/7	2015	RGB	Urban	✔️	✔️	✔️		GPS
	SPED	2017	RGB	Outdoor	✔️	✔️			-
Omni-directional	New College Vision and Laser	2009	Gray.	Outdoor	slight	✔️	✔️		GPS, IMU, LiDAR
	MOLP	2018	Gray./D	Outdoor	✔️		✔️		GPS
	NCLT	2016	RGB	Outdoor	✔️	✔️	✔️		GPS, LiDAR
Aerial/UAV	Shopping Street 1/2	2018	Gray.	Urban	slight	✔️	✔️		-
	EuRoC	2016	Gray.	Indoor		✔️	✔️		IMU
Underwater	UWSim	2016	RGB	Under-water			✔️		GPS
Range sensors	MulRan	2020	3D Point clouds	Urban	✔️		✔️		LiDAR, RADAR

Reference:

[1] Cummins, M. & Newman, P. FAB-MAP: Probabilistic Localization and Mapping in the Space of Appearance. The International Journal of Robotics Research, 2008, 27, 647-665

[2] M. Smith, I. Baldwin, W. Churchill, R. Paul, P. Newman, The new college vi- sion and laser data set, Int. J. Rob. Res. 28 (5) (2009) 595–599, doi: 10.1177/ 0278364909103911 .

• TODO ...

---

Citations to this work:

@article{ZHANG2020107760,
title = "Visual place recognition: A survey from deep learning perspective",
journal = "Pattern Recognition",
pages = "107760",
year = "2020",
issn = "0031-3203",
doi = "https://doi.org/10.1016/j.patcog.2020.107760",
url = "http://www.sciencedirect.com/science/article/pii/S003132032030563X",
author = "Xiwu Zhang and Lei Wang and Yan Su",
}