Sea-Land Segmentation of Remote Sensing Images Based on SDW-UNet

作     者：Tianyu Liu Pengyu Liu Xiaowei Jia Shanji Chen Ying Ma Qian Gao 

作者机构：The Information DepartmentBeijing University of TechnologyBeijing 100124China School of Physics and Electronic Information EngineeringQinghai Minzu UniversityXining810000China Advanced Information Network Beijing LaboratoryBeijing100124China Computational Intelligence and Intelligent Systems Beijing key LaboratoryBeijing100124China Department of Computer ScienceUniversity of Pittsburgh15260USA 

出 版 物：《Computer Systems Science & Engineering》 (计算机系统科学与工程（英文）)

年 卷 期：2023年第45卷第5期

页      面：1033-1045页

核心收录：

学科分类：0810[工学-信息与通信工程] 08[工学] 081002[工学-信号与信息处理] 

基　　金：This paper is supported by the following funds:The National Key Research and Development Program of China(2018YFF01010100) The Beijing Natural Science Foundation(4212001) Basic Research Program of Qinghai Province under Grants No.2021-ZJ-704 Advanced information network Beijing laboratory(PXM2019_014204_500029) 

主　　题：Sea-land segmentation UNet depth-wise separable convolution squeeze-excitation position encoding 

摘      要：Image segmentation of sea-land remote sensing images is of great importance for downstream applications including shoreline extraction,the monitoring of near-shore marine environment,and near-shore target *** mitigate large number of parameters and improve the segmentation accuracy,we propose a new Squeeze-Depth-Wise UNet(SDW-UNet)deep learning model for sea-land remote sensing image *** proposed SDW-UNet model leverages the squeeze-excitation and depth-wise separable convolution to construct new convolution modules,which enhance the model capacity in combining multiple channels and reduces the model *** further explore the effect of position-encoded information in NLP(Natural Language Processing)domain on sea-land segmentation *** have conducted extensive experiments to compare the proposed network with the mainstream segmentation network in terms of accuracy,the number of parameters and the time cost for *** test results on remote sensing data sets of Guam,Okinawa,Taiwan China,San Diego,and Diego Garcia demonstrate the effectiveness of SDW-UNet in recognizing different types of sea-land areas with a smaller number of parameters,reduces prediction time cost and improves performance over other mainstream segmentation *** also show that the position encoding can further improve the accuracy of model segmentation.