Landslide detection from an open satellite imagery and digital elevation model dataset using attention boosted convolutional neural networks
Citation
Ji, S., Yu, D., Shen, C., Li, W., Xu, Q. (2020). Landslide detection from an open satellite imagery and digital elevation model dataset using attention boosted convolutional neural networks. Landslides, 17: 1337-1352. Link to paper
Abstract
Attention mechanisms originated from the human visual system are developed for boosting the CNN to extract more distinctive feature representations of landslides from backgrounds. The implementation of the attention module in CNN models in remote sensing image processing can improve the global context modeling and feature detection of the model. The Bijie landslide dataset, an open source satellite imagery dataset, includes optical satellite images, shapefiles that depict landslide borders, labels, and digital elevation models. The dataset includes 770 RGB images of landslides and 2003 non-landslide images at a resolution of 0.8 m. Results demonstrate that attention-boosted models achieve superior performance in landslide detection tasks compared to standard CNN architectures.