Earthquake-induced landslide recognition using high-resolution remote sensing images

Abstract

Earthquake-induced landslides are the most common geological disasters caused by large seismic activities in mountainous areas, and they are known for their suddenness, destructiveness, and extensive distribution range. These landslides often result in severe casualties and economic losses. Currently, regional earthquake-induced landslides are mainly obtained by visual interpretation and computer data extraction from remote sensing images. These methods are objective, time-consuming, and low in precision. Thus, they cannot address the requirement of practically conducting emergency surveys and disaster evaluations after earthquakes. In this study, with the main data source of high-resolution remote sensing images from ZY-3 and GF-1, as well as the study area of the Wenchuan earthquake region, objects of multilevel landslides were established using the multi-scale optimum partition method based on an in-depth analysis of seismic landslide features. A recognition rule set of multi-dimensional landslides was also built by combining topographic and image features, such as spectrum, texture, and geometry. Additionally, recognition models for landslide stratification were proposed based on the recognition models of high-resolution images and an understanding of the scenes. Through all of the preceding efforts mentioned, the spatial distribution of the seismic landslide, as well as the sliding source, transport, and depositional areas, can be identified. The analysis results of the experimental area showed a minimum recognition accuracy of 81.89%, with the depositional zone of landslides being the easiest zone to recognize, and the established method can be generalized. These findings may provide technical support for post-earthquake emergency investigations and further promote the application of high-resolution remote sensing data from Chinese satellites for landslides recognition.