TeLIP: Terrain-Based Landslide Initiation Position Model

This dataset is designed to support the training and evaluation of the <b>TeLIP (Terrain-based Landslide Initiation Position Model)</b>, a deep learning model aimed at predicting landslide initiation positions based on topographic and geologic features of hillslopes. The dataset includes the following three types of input data, which are essential for the model's operation:<b>Geological Data</b>: Contains various predictors such as relative height, elevation, slope, soil thickness, profile curvature, plan curvature, TWI (Topographic Wetness Index), and position index. These features are used to capture the spatial characteristics of the terrain.<b>Landslide Position Data</b>: The labels in this dataset indicate whether a given position along the slope is the initiation point of a landslide. These binary classification labels are essential for training the model to predict the landslide initiation positions.<b>Slope Length Data</b>: Provides the total length of each slope segment, which is necessary for evaluating prediction errors in terms of the distance between predicted and true landslide positions.Users can create their corresponding training datasets based on the following data format according to their landslide research objectives.The source code of the TeLIP framework is available on GitHub at https://github.com/Wujunyu0305/TeLIP