The automatic identification of loess landslide is an important work, which can effectively help the risk management of landslide disaster. Traditional landslide identification methods mainly rely on manual operation, and the efficiency is low. Therefore, this paper explores deep learning method and uses remote sensing image to automatically identify loess landslide. Using Google maps open source dataset and using deep learning model to achieve automatic detection of landslides. The experimental results show that the proposed method can significantly improve the accuracy and recall rate, which provides strong support for the automatic recognition of loess landslides.
维普
万方数据