The point cloud is obtained using Ouster, Aeva, Livox, and Velodyne. All scense are acquired at same location, varying scene apperances
Recent Update
2024 / 03: HeLiPR dataset is accepted in IJRR Journal.
2024 / 01: We added 3 sequences KAIST06, DCC06 and Riverside06. These sequences can be utilized for validating temporal discrepancy with previous Dajeon sequences.
2023 / 11 : We update the description for system to introduce the sensor configuration in detail.
2023 / 09 : Our paper is submitted to the International Journal of Robotics Research.
2023 / 09 : The similar trajectory with MulRan is added for long-term place recognition (4 year differences). Furthermore, global pose (utm) for each LiDAR is updated.
2023 / 09 : Our paper (arxiv version) is uploaded.
2023 / 09 : Our paper is accepted to the Novel Sensors workshop at IROS 2023. This dataset is selected as invited spotlight presentation.
2023 / 08 : The groundtruths of rotation part for all seqeunces are updated correctly.
2023 / 08 : The HeLiPR dataset homepage is firstly opened.
Contact
Official : rpmsnu@gmail.com, helipr.dataset@gmail.com
Maintainer : Minwoo Jung (moonshot@snu.ac.kr) [CV] [LinkedIn] [GitHub]