├── README.md ├── calibrations ├── extrinsic.yaml ├── param_camera.yaml └── param_imu.yaml └── data ├── bdcalib.jpg ├── bigcar2.jpg └── ground_truth.txt /README.md: -------------------------------------------------------------------------------- 1 | # SJTU_GVI 2 | A GNSS-Visual-IMU benchmark Dataset for SLAM 3 | 4 | 5 | 6 | 7 | [![Author](https://img.shields.io/badge/Author-Jie%20Yin-blue)](https://sjtuyinjie.github.io/) 8 | [![Paper](https://img.shields.io/badge/Paper-M2CVIO-yellow)](https://link.springer.com/article/10.1186/s43020-023-00102-9) 9 | [![License](https://img.shields.io/badge/License-MIT-cyan)]() 10 | [![News](https://img.shields.io/badge/News-orange)](https://mp.weixin.qq.com/s/E2URbAKHCYq37XRpDd6cuQ) 11 | 12 | This is part of the dataset for tests in paper [M2C-GVIO](https://satellite-navigation.springeropen.com/articles/10.1186/s43020-023-00102-9) 13 | Different from M2DGR, this dataset has following features: 14 | 15 | 1.More light-weight and easy for downloading. 16 | 17 | 2.Recoreded on a real car with high speed. 18 | 19 | 3.GNSS raw measurements are captured by a Ublox ZED-F9P receiver, which facilitate the GNSS-SLAM research. 20 | 21 | Please give us a star if this project is helpful to your research. Thank you! If you use M2DGR in an academic work, please cite: 22 | 23 | ~~~ 24 | @ARTICLE{9664374, 25 | author={Yin, Jie and Li, Ang and Li, Tao and Yu, Wenxian and Zou, Danping}, 26 | journal={IEEE Robotics and Automation Letters}, 27 | title={M2DGR: A Multi-sensor and Multi-scenario SLAM Dataset for Ground Robots}, 28 | year={2021}, 29 | volume={}, 30 | number={}, 31 | pages={1-1}, 32 | doi={10.1109/LRA.2021.3138527}} 33 | 34 | @article{hua2023m2c, 35 | title={M2C-GVIO: motion manifold constraint aided GNSS-visual-inertial odometry for ground vehicles}, 36 | author={Hua, Tong and Pei, Ling and Li, Tao and Yin, Jie and Liu, Guoqing and Yu, Wenxian}, 37 | journal={Satellite Navigation}, 38 | volume={4}, 39 | number={1}, 40 | pages={1--15}, 41 | year={2023}, 42 | publisher={SpringerOpen} 43 | } 44 | ~~~ 45 | 46 |
47 | 48 | 49 |
50 | 51 |

Figure 1. A picture of our acquisition platform.

52 | 53 | 54 | 55 |
56 | 57 | 58 |
59 |

Figure 2. We were calibrating the extrinsics.

60 | 61 | ## 1.Sequence 62 | The extraction code is "yj66" 63 | 64 | 65 | Sequence|Collection Date|Total Size|Duration|Rosbag 66 | --|:--|:--:|--:|--: 67 | Seq1|2021-12-23|2.37G|349s|[Rosbag](https://pan.baidu.com/s/1qoFsvOUyJCf7xi6KHBBoCQ) 68 | Seq2|2021-12-23|921M|109s|[Rosbag](https://pan.baidu.com/s/15usabuNPmlmC_S4cQsGIYg) 69 | Seq3|2021-12-23|1.19G|146s|[Rosbag](https://pan.baidu.com/s/1AfoDweqbo89PDl6ef0mYZA) 70 | Seq4|2021-12-23|850M|112s|[Rosbag](https://pan.baidu.com/s/1uc6RLjXhjs15g7Czg_JTDA) 71 | Seq5|2021-12-23|1.26G|159s|[Rosbag](https://pan.baidu.com/s/1vlE16qNDnV4C2i-cfg6NVQ) 72 | ## 2.Calibration 73 | The camera intrinsic and cam-imu extrinsics calibration results is given in [Link](https://github.com/sjtuyinjie/SJTU_GVI/tree/main/calibrations) 74 | ## 3.Ground Truth 75 | The ground truth of the datasets is given in [Link](https://github.com/sjtuyinjie/SJTU_GVI/tree/main/data) 76 | 77 | 78 | 79 | ## 4.ACKNOWLEGEMENT 80 | Authors express our appreciation for the support of Shanghai West Hongqiao Navigation Technology Co., LTD. 81 | -------------------------------------------------------------------------------- /calibrations/extrinsic.yaml: -------------------------------------------------------------------------------- 1 | %YAML:1.0 2 | 3 | # extrinsic parameters for SJTU GVI. 4 | 5 | T_ib (body to imu): 6 | [ 1., 0., 0., 0., 7 | 0., 1., 0., 0., 8 | 0., 0., 1., 0., 9 | 0., 0., 0., 1. ] 10 | 11 | T_ci (imu0 to cam0): 12 | [ 0.99964049, 0.02680926, -0.00040322, 0.04313484, 13 | -0.02680946, 0.99964044, -0.00048997, 0.00735024, 14 | 0.00038994, 0.00050060, 0.99999980, 0.03880331, 15 | 0.00000000, 0.00000000, 0.00000000, 1.00000000 ] 16 | 17 | 18 | timeshift cam0 to imu0 (s): 0.0138260221117 19 | -------------------------------------------------------------------------------- /calibrations/param_camera.yaml: -------------------------------------------------------------------------------- 1 | %YAML:1.0 2 | --- 3 | 4 | Camera model: pinhole 5 | 6 | Image width: 640 7 | Image height: 480 8 | 9 | Focal length: [941.012161218307, 938.194870227585] 10 | Principal point: [636.816098847571, 357.728952908842] 11 | 12 | Distortion model: radtan 13 | Distortion coefficients: [0.165049318972363, -0.281973486275739, 0.0, 0.0] 14 | -------------------------------------------------------------------------------- /calibrations/param_imu.yaml: -------------------------------------------------------------------------------- 1 | %YAML:1.0 2 | --- 3 | 4 | type: imu 5 | name: d435i 6 | 7 | Update rate: 200.0 8 | 9 | Accelerometer: 10 | Noise density: 0.0268487616106 11 | Noise density (discrete): 0.379698828027 12 | Random walk: 0.00085216274964 13 | Gyroscope: 14 | Noise density: 0.00247107870753 15 | Noise density (discrete): 0.0349463302188 16 | Random walk: 0.0000179631459052 17 | -------------------------------------------------------------------------------- /data/bdcalib.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/sjtuyinjie/SJTU_GVI/c54f24b8f581713955776649613ed008301f34ba/data/bdcalib.jpg -------------------------------------------------------------------------------- /data/bigcar2.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/sjtuyinjie/SJTU_GVI/c54f24b8f581713955776649613ed008301f34ba/data/bigcar2.jpg --------------------------------------------------------------------------------