PASTIS Large-Scale Panoramic SITS Dataset

PASTIS, short for Panoptic Segmentation of satellite image TImes Series, is the first open SITS dataset with panoramic annotations. The dataset consists of 2,433 multispectral image sequences of 10×128×128 shape. Each sequence contains 38 to 61 observation points between September 2018 and November 2019, totaling more than 2 billion pixels. The acquisition interval is uneven, with an average of 5 days.

Each pixel in the dataset has semantic annotations, with a total of 19 categories, including 18 crop types and 1 background class. The annotations are taken from a French public land parcel recognition system with an annotation accuracy of more than 98%. For cross-validation, the dataset was divided into 5 folds.