DeepOWT: a global offshore wind turbine data set derived with deep learning from Sentinel-1 data

Abstract. Offshore wind energy is at the advent of a massive global expansion. To investigate development offshore sector, optimal farm locations, or impact projects, freely accessible spatiotemporal data set infrastructure necessary. With free and direct access to such data, it more likely that all stakeholders who operate in marine coastal environments will become involved upcoming expansion farms. end, we introduce DeepOWT (Deep-learning-derived Wind Turbines) (available https://doi.org/10.5281/zenodo.5933967, Hoeser Kuenzer, 2022b), which provides 9941 locations along with their deployment stages on scale. based Earth observation from Sentinel-1 radar mission. The were derived by applying deep-learning-based object detection two cascading convolutional neural networks (CNNs) search entire archive successive CNNs have previously been optimised solely synthetic training examples detect infrastructures real-world imagery. subsequent temporal analysis signal detected reports each quarterly frequency July 2016 until June 2021. information compiled ready-to-use geographic system (GIS) format make usability as possible.