Earthquake-triggered landslides (ETLs) are characterized by their extensive occurrence,having wide distributions. The conventional human-computer interaction extraction method is often time-consuming and labor-intensive, failing to meet the demands of disaster emergency response. There is a pressing need for a swift detection of ETLs. In this study, we introduce an earthquake triggered landslides extraction method (MAST) combining morphological analysis edge recognition with Swin-Transformer deep learning model which is specifically designed for landslide extraction. The MAST model adopts a hierarchical construction approach akin to convolution neural networks, aiding in tasks such as target detection and semantic segmentation. To enhance the accuracy of landslide edge extraction, we incorporate an edge recognition algorithm based on morphological analysis into the MAST model. This algorithm leverages morphological operations to extract the features of landslide boundaries. It effectively addresses issues such as discretization and irregularization of the extracted landslide boundaries, leading to more precise delineation of landslide boundaries.Drawing on UAV data collected from Wan Dong Village, De Tou Town, Sichuan Luding, China, during the 2022 Ms 6.8 Luding Earthquake, we conduct automated extraction of ETLs utilizing the MAST model. Experimental results demonstrate the superior performance of the MAST model compared to the traditional full convolution neural network model (FCN)and normal Swin-Transformer model (SWT). The MAST model exhibits enhanced most value in landslide extraction. Notably, it demonstrates a significant advantage in boundary extraction. Employing the Boundary IoU metric to evaluate the accuracy of ETLs extraction, the MAST model outperforms the SWT and FCN models at various distances.

The magnitude 6.8 Luding earthquake that occurred on 5 September 2022, triggered multiple large-scale landslides and caused a heavy loss of life and property. The investigation of earthquake-triggered landslides (ETLs) facilitates earthquake disaster assessments, rescue, reconstruction, and other post-disaster recovery efforts. Therefore, it is important to obtain landslide inventories in a timely manner. At present, landslide detection is mainly conducted manually, which is time-consuming and laborious, while a machine-assisted approach helps improve the efficiency and accuracy of landslide detection. This study uses a fully convolutional neural network algorithm with the Adam optimizer to automatically interpret the aerial and satellite data of landslides. However, due to the different resolutions of the remote sensing images, the detected landslides vary in boundary and quantity. In this study, we conducted an assessment in the study area of Wandong village in the earthquake-affected area of Luding. UAV images, GF-6 satellite images, and Landsat 8 satellite images, with a resolution of 0.2 m, 2 m, and 15 m, respectively, were selected to detect ETLs. Then, the accuracy of the results was compared and verified with visual detection results and field survey data. The study indicates that as the resolution decreases, the accuracy of landslide detection also decreases. The overall landslide area detection rate of UAV imagery can reach 82.17%, while that of GF-6 and Landsat 8 imagery is only 52.26% and 48.71%. The landslide quantity detection rate of UAV imagery can reach 99.07%, while that of GF-6 and Landsat 8 images is only 48.71% and 61.05%. In addition, for each landslide detected, little difference is found in large-scale landslides, and it becomes more difficult to correctly detect small-scale landslides as the resolution decreases. For example, landslides under 100 m² could not be detected from a Landsat 8 satellite image.