Remote Sensing Image Segmentation of Mariculture Cage Using Ensemble Learning Strategy

In harbour areas, the irrational layout and high density of mariculture cages can lead to a dramatic deterioration culture’s ecology. Therefore, it is important analyze regulate distribution using intelligent analysis based on deep learning. We propose remote sensing image segmentation method Swin Transformer ensemble learning strategy. Firstly, we collect multiple images annotate them, while data expansion techniques construct dataset cages. Secondly, used as backbone network extract features A strategy alternating local attention module global for model training, which has benefit reducing computation exchanging information. Then, improve accuracy cage segmentation. carry out quantitative qualitative analyses at ports Li’an, Xincun Potou in Hainan Province, China. The results show that our proposed scheme significant performance improvement compared other models. particular, mIoU reaches 82.34% pixel 99.71%.