DeepOWT: A global offshore wind turbine data set derived with deep learning from Sentinel-1 data
Thorsten Hoeser, Stefanie Feuerstein, Claudia Kuenzer
Abstract
Abstract. Offshore wind energy is at the advent of a massive global expansion. Driven by carbon neutral alternatives for energy generation, offshore wind energy receives growing attention as a renewable energy source. Despite the large amount of unused wind energy capacities worldwide, offshore wind farms have to be integrated into already intensively used maritime economic areas. The optimal choice of offshore wind farm locations is as crucial as compatibility with other stakeholders while minimising ecological impacts. Thus, a spatiotemporal data set for offshore wind turbine deployment is necessary to involve all stakeholders and exchange knowledge during the upcoming massive expansion of offshore wind farms. To that end, we introduce the DeepOWT data set (global offshore wind turbines derived with deep learning; available at: https://doi.org/10.5281/zenodo.5933967 (Hoeser and Kuenzer, 2022), which provides 9,941 locations of offshore wind energy infrastructure along with their deployment stages on a global scale. DeepOWT is based on freely accessible Earth observation data from 2016 until 2021. The locations were derived from radar imagery of the Sentinel-1 mission by applying deep learning based object detection, trained on synthetic training examples. The entire deployment process is reported in a quarterly frequency and spatially contextualised for each single wind turbine location in a ready to use GIS format. Therewith, the DeepOWT data set can directly be used to enable spatial planning, environmental investigations and to optimise location decisions and the deployment process.