SimuShips - A High Resolution Simulation Dataset for Ship Detection with Precise Annotations

SimuShips - A High Resolution Simulation Dataset for Ship Detection with Precise Annotations Availability of domain-specific datasets is a challenge for object detection. In particular, conducting onsite experiments is risky and time-consuming for the maritime domain. It is not always possible to capture challenging situations and incorporate variations in the environment. Therefore, we present a simulation-based dataset for maritime environments consisting of 9471 high-resolution (1920x1080) images resembling the real world with precise annotations. The dataset was acquired using a simulation tool, AILiveSim [17]. AILiveSim is a 3D simulation platform for targeted scalable development and integration of autonomous systems through digital twins. The tool provides a realistic 3D model of the route of the watercraft extended from the city of Turku to Ruisalo in South-West Finland. Our dataset incorporates diversity of objects, weather, illumination, visible proportion and scale in images that resemble the real world. Please cite the following publication when using the dataset: M. Raza, H. Prokopova, S. Huseynzade, S. Azimi and S. Lafond, "SimuShips - A High Resolution Simulation Dataset for Ship Detection with Precise Annotations," OCEANS 2022, Hampton Roads, Hampton Roads, VA, USA, 2022, pp. 1-5, doi: 10.1109/OCEANS47191.2022.9977182. The publication is available at: https://doi.org/10.1109/OCEANS47191.2022.9977182 A preprint version of the publication is available at https://arxiv.org/abs/2211.0523