This zipfile contains data to accompany SlideDetect: Spatio-Temporal Landslide Detection Using a Three Dimensional Convolutional Neural Network This repository contains processed data used to train a 3D convolutional neural network to detect landslides in satellite imagery, as follows: - images.npy: processed 4-band RGBA imagery for the study area - images_land_cover.npy: processed 13-band imagery for study area, with the RGBA imagery for the first 4 bands, and 1-hot encoding of land cover as the final 9 bands. - image_stack_indices.npy: list of indices of 4-D temporal image stacks used in model training. Indices are stored in the format [xstart, xstop, ystart, ystop, zstart, zstop] - labels.npy: processed ground-truth data used to train the model. Array is the 1-band image stack in same shape as images.npy, but with a binary label indicating whether a new landslide occurs at that location and time stamp.