CAV-PR (CERTH Autonomous Vehicle - Place Recognition) is a public dataset which contains sychronized sequences of RGB image, 3D Lidar, IMU and GPS-RTK measurements. It composed of two sessions and each one of them contain revisits from both same and reverse viewpoints. This is the official repository of the paper:
Loop closure detection with 3D LiDAR Data for Extreme Viewpoint Changes
Contents: Please find and download the dataset with its calibration file.
https://www.kaggle.com/dalexiou/certh-robot-car-place-recognition-dataset
Format
Each session is saved in ROSBAG file.
Calib.txt includes the neccessary calibrations between the sensors.
Citation:
D. Alexiou, K. Tsiakas, I. Kostavelis, D. Giakoumis, A. Gasteratos and D. Tzovaras, "Loop-Closure Detection with 3D LiDAR Data for Extreme Viewpoint Changes," 2022 26th International Conference on Methods and Models in Automation and Robotics (MMAR), 2022, pp. 29-34, doi: 10.1109/MMAR55195.2022.9874344.
Contact:
For information/questions about the paper or the dataset, please contact Dimitrios Alexiou