FastICENet: A real-time and accurate semantic segmentation model for aerial remote sensing river ice image

•A two-branch structure and a learnable upsampling strategy DUpsampling is adopted to construct an accurate ice semantic segmentation network.•A new Downsampling module and a dense connection block in the context branch is designed to construct a real-time ice semantic segmentation network.•Two river ice segmentation datasets NWPU_YRCC_EX and NWPU_YRCC2 are enlarged and relabeled.•The proposed method can obtain better results in practical application.