Study on Earthquake-triggered Landslides Recognition in Cloudy Areas

Earthquakes can trigger numerous and widely distributed landslides. Rapid mapping of landslides with high accuracy is crucial for understanding their spatial patterns, and assessing landslide hazards and risks. Visual interpretation or automatic recognition are conventional methods for earthquake-triggered landslide mapping. Because of the influence of cloud, the quality of high-resolution optical remote sensing images is low, which limit investigation and assessment of earthquake-triggered landslides. Based on the Google Earth Engine platform, this work proposed the method for masking cloud and mosaicing cloud-free images using Sentinel-2 remote sensing images and cloud probability dataset. The proposed method can improve the quality of post-earthquake images. Then, an object-oriented model for earthquake-triggered landslide recognition is constructed based on multiscale segmentation and many features including spectral features, vegetation index and spatial proximity relationship between different classes. This work takes 2021 Haiti Nippes earthquake as study area. The results show that this method benefits earthquake-triggered landslides detection in the cloudy area and can provide technical support for post-disaster emergency investigation and assessment.