Sinkholes are the most abundant surface features in karst areas worldwide. Understanding sinkhole occurrences and characteristics is critical for studying karst aquifers and mitigating sinkhole-related hazards. Most sinkholes appear on the land surface as depressions or cover collapses and are commonly mapped from elevation data, such as digital elevation models (DEMs). Existing methods for identifying sinkholes from DEMs often require two steps: locating surface depressions and separating sinkholes from non-sinkhole depressions. In this study, we explored deep learning to directly identify sinkholes from DEM data and aerial imagery. A key contribution of our study is an evaluation of various ways of integrating these two types of raster data. We used an image segmentation model, U-Net, to locate sinkholes. We trained separate U-Net models based on four input images of elevation data: a DEM image, a slope image, a DEM gradient image, and a DEM-shaded relief image. Three normalization techniques (Global, Gaussian, and Instance) were applied to improve the model performance. Model results suggest that deep learning is a viable method to identify sinkholes directly from the images of elevation data. In particular, DEM gradient data provided the best input for U-net image segmentation models to locate sinkholes. The model using the DEM gradient image with Gaussian normalization achieved the best performance with a sinkhole intersection-over-union (IoU) of 45.38% on the unseen test set. Aerial images, however, were not useful in training deep learning models for sinkholes as the models using an aerial image as input achieved sinkhole IoUs below 3%.
Digital Object Identifier (DOI)
This material is based upon work supported by the National Science Foundation under Grant No (IIS-1553116). The second author is supported by the National Science Foundation Grant No (EAR-1933779).
Model source code, installation instructions, and scripts for training and inference are in Github at https://mvrl.github.io/SinkSeg/. The image data set used in the model is deposited at https://doi.org/10.5281/zenodo.5789436. Data sources used to derive the image data set are available in these in-text data citation references: aerial imagery from the National Agriculture Imagery Program (KyFromAbove, n.d.), (public domain); digital elevation model derived from LiDAR data (KyFromAbove, n.d.), (public domain); binary label image derived from Kentucky LiDAR-derived sinkholes (Kentucky Geological Survey, n.d.), (public domain); digital elevation model for Missouri from Missouri Spatial Data Information Service (Missouri Spatial Data Information Service, n.d.), (public domain); and sinkhole data for Greene County, Missouri from City of Springfield, Missouri (City of Springfield, Missouri, n.d.), (public domain).
Published in Earth and Space Science, v. 9, issue 2, e2021EA002195. © 2022 The Authors This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.