海军航空大学学报2024，Vol.39Issue(5)：535-545,586,12.DOI:10.7682/j.issn.2097-1427.2024.05.003
深度学习赋能的高光谱图像分类研究进展
Research Progress in Deep Learning-Enabled Hyperspectral Image Classification
摘要
Abstract
With the development of hyperspectral imaging technology,hyperspectral image classification has become a re-search field of great interest.Based on extensive research,the hyperspectral image classification methods based on deep learning are organized comprehensively,mainly covering deep networks,recurrent networks and self-attention networks.Subsequently,several representative methods are discussed in depth,and the advantages and shortcomings of these meth-ods are explored in detail,aiming to provide a clearer and more comprehensive picture of the current status of hyperspec-tral image classification methods.A comprehensive overview of hyperspectral image classification methods is provided and an in-depth study of various methods is conducted,the qualitative and quantitative evaluation results of different methods are analyzed,and an outlook on the future development is provided.Sorting through existing research not only helps to promote the further development of hyperspectral remote sensing technology,but also highlights the unique ad-vantages of hyperspectral image classification methods in aerospace and other fields,which is of great significance to im-prove the interpretation accuracy and practical application value of remote sensing data.关键词
高光谱图像分类/深度网络/循环网络/自注意力网络Key words
hyperspectral image classification/deep networks/recurrent networks/self-attention networks分类
信息技术与安全科学引用本文复制引用
白林锋,陈增俊,周玲,张妍妍,路凯,张卫东..深度学习赋能的高光谱图像分类研究进展[J].海军航空大学学报,2024,39(5):535-545,586,12.基金项目
河南省科技攻关项目(242102210075、232102210018、242102211059) （242102210075、232102210018、242102211059）
河南省重点研发计划(241111211800) （241111211800）
河南省自然科学青年基金(232300420428) （232300420428）
河南省教师教育课程改革研究(2024-JSJYYB-099) （2024-JSJYYB-099）
