Semi-automatic mapping of large shallow landslides using high-resolution remote sensing imagery

A semi-automatic and hierarchical detection method using only post-event high-resolution (HR) remote sensing imagery for recent shallow landslides that are triggered by an earthquake or a rainfall event, were introduced in this study. Firstly, candidate landslide regions are obtained by extracting objects in light- tone using HR images. Then, shape parameters are employed elimination roads and building in light-tone, and text parameters are considered to remove farmlands from the candidate regions. The proposed method achieves average accuracies of approximately 77% and 68%, in terms of landslide number and area, respectively.