An Open Hyperspectral Dataset with Sea-Land-Cloud Ground-Truth from the HYPSO-1 Satellite

Datasets from airborne sensors like AVIRIS [2], ROSIS [3], and HYDICE [4] contain Hyperspectral Imaging, employed in satellites for space remote labeled images with diverse land cover categories like urban sensing, like HYPSO-1, faces constraints due to few and agricultural areas. Additionally, other popular datasets labeled data sets, affecting the training of AI models demanding such as the Kennedy Space Center and Jasper Ridge have limited these ground-truth annotations. In this work, we water coverage due to the capture over small geographic introduce The HYPSO-1 Sea-Land-Cloud-Labeled Dataset, extents using airborne platforms. Despite the widespread use an open dataset with 200 diverse hyperspectral images from of these datasets in HSI classification, each set consists of a the HYPSO-1 mission, available in both raw and calibrated single labeled image, inadequate for training emerging dataintensive forms for scientific research in Earth observation.