Continual Learning Based Ship Detection in Multiple Weathers
Baosheng Li, Jishui Han, Zhenliang Qi, Liqiang Gao, Ruijie Duan, Tongtong Wang, Peng Yan, Ran Song, Xiaolei Li · 2022
Ship detection in complex weathers is an important application of object detection approaches in real life. In recent years, ship detection approaches for single weather have gradually matured. However, there are few ship detection approaches for multiple weathers, and traditional ship detection approaches are only applicable to a specific weather and have poor generalization to other weathers. In this paper, we propose a continual learning based approach for ship detection to solve the task of ship detection in various weather conditions, while enabling the model to be capable of expanding to new weather conditions. We also present a ship detection dataset under three types of weather, including rain, fog, and snow, using weather effect simulation methods.