Multimodal Vision Research Laboratory

MVRL

Research Area: Geohazards and Terrain

How can we map hazardous terrain and environmental disturbances from elevation, LiDAR, and overhead imagery? We develop segmentation and fusion methods for sinkholes, landslides, and post-disturbance landscape change. Recent work includes attention-enhanced fusion of elevation and aerial imagery for sinkhole segmentation, deep learning assessment of landslide mapping from LiDAR-based elevation data, and mapping post-fire permafrost degradation in arctic tundra. This area complements our geospatial AI methods with applications in civil engineering, karst science, and environmental monitoring where reliable terrain semantics matter.

All Publications

  1. Zhu J, Cunningham D, Jacobs N. 2026. Attention-Enhanced Multimodal Fusion of Elevation and Aerial Imagery for Sinkhole Segmentation. In: Multidisciplinary Conference on Sinkholes and the Engineering and Environmental Impacts of Karst.
  2. Koch HJ, Zhu J, Bhatt S, Jacobs N, Dortch JM. 2026. Assessing deep learning segmentation models for mapping landslides from lidar-based elevation data. DOI: 10.22541/essoar.177265426.62613754/v1.
  3. Badzioch R, Jacobs N, Rastetter EB, Rocha AV. 2025. Mapping Post-Fire Polygonal Ice Wedge Degradation in Arctic Tundra Using High Resolution Satellite Imagery and Computer Learning. In: American Geophysical Union (AGU) Fall Meeting Abstracts.
  4. Rafique MU, Zhu J, Jacobs N. 2021. Automatic Segmentation of Sinkholes Using a Convolutional Neural Network. Earth and Space Science:19. DOI: 10.1002/essoar.10509794.1.
  5. Zhu J, Nolte A, Jacobs N, Ye M. 2020. Machine Learning in Identifying Karst Sinkholes from LiDAR-Derived Topographic Depressions in the Bluegrass Region of Kentucky. Journal of Hydrology. DOI: 10.1016/j.jhydrol.2020.125049.
  6. Zhu J, Nolte AM, Jacobs N, Ye M. 2019. Incorporating Machine Learning with LiDAR for Delineating Sinkholes. In: Kentucky Water Resources Annual Symposium.