Symphony Lake Dataset is available here.

The Symphony Lake

The goal of this project is to monitor the shore of the Lake Symphonie in Metz, France with high-frequency over a significant time scale, using our Kingfisher platform. From these dataset, it will be possible to build an history of geo-referenced changes. This project is a stepping stone towards the deployment of the Kingfisher in more complex or more relevant missions. This will let us develop the autonomy component required for any real mission, test them in real scenarios and make sure their robustness level is satisfactory.

Here is a summary of the main research approach by Shane Griffith.

Fast summary of a 1-hour lake survey.

3D reconstruction of the lake shore using latest results from Xiaolong Wu.

Hidden heron.

Typical lake shore.

Sometimes a curious swan comes to see us.

Illustration of the lakeshore dataset we are collecting.

Automatically aligned timelapse accross seasons (S. Griffith).

Image pair selections from Multi-Session ICP Mapping.

Bibliography

  1. Griffith, S., Dellaert, F., & Pradalier, C. (2020). Transforming multiple visual surveys of a natural environment into time-lapses. The International Journal of Robotics Research, 39(1), 100–126. https://hal.archives-ouvertes.fr/hal-02278909
  2. Griffith, S., Chahine, G., & Pradalier, C. (2017). Symphony lake dataset. The International Journal of Robotics Research, 36(11), 1151–1158. PDF
  3. Griffith, S., & Pradalier, C. (2017). Survey registration for long-term natural environment monitoring. Journal of Field Robotics, 34(1), 188–208. PDF
  4. Pradalier, C., & Griffith, S. (2017). Long-term monitoring of a natural environment using an automated data acquisition system. ILTER Network.
  5. Griffith, S., & Pradalier, C. (2016). Reprojection Flow for Image Registration Across Seasons. BMVC. PDF
  6. Griffith, S., & Pradalier, C. Towards Reprojection Flow for Image Registration Across Seasons. PDF
  7. Griffith, S., Dellaert, F., & Pradalier, C. (2015). Robot-enabled lakeshore monitoring using visual SLAM and SIFT flow. RSS Workshop on Multi-View Geometry in Robotics. PDF
  8. Griffith, S., & Pradalier, C. (2016). A spatially and temporally scalable approach for long-term lakeshore monitoring. Field and Service Robotics: Results of the 10th International Conference, 3–16. PDF
  9. Griffith, S., Drews, P., & Pradalier, C. (2014). Towards autonomous lakeshore monitoring. Experimental Robotics: The 14th International Symposium on Experimental Robotics, 545–557. PDF