摘要
多特征决策融合是模式识别中常用的手段之一,可以有效结合不同类特征的优点从而提高分类性能。论文分析了三类典型的多特征决策融合算法并讨论了它们在无人机合成孔径雷达(SAR)目标识别中的运用效能。它们包括并行决策融合、级联决策融合以及联合决策融合。实验中,以MSTAR数据集为基础分类了这三类多特征决策融合方法对于SAR目标识别性能的影响。
Multi-feature decision fusion is a usual method in pattern recognition field,which could effectively combine the ad⁃vantages of different kinds of features.In this paper,three typical multi-feature decision fusion methods are analyzed and their ap⁃plications to target recognition of unmanned aerial vehicle(UAV)synthetic aperture radar(SAR)images are discussed.The three methods are parallel decision fusion,hierarchical decision and joint decision fusion.In the experiments,the MSTAR dataset is em⁃ployed to investigate the effects of three multi-feature decision fusion methods on the performance of SAR target recognition.
作者
马梓元
龚华军
王新华
刘禹
MA Ziyuan;GONG Huajun;WANG Xinhua;LIU Yu(School of Automation,Nanjing University of Aeronautics and Astronautics,Nanjing 211100)
出处
《舰船电子工程》
2020年第12期96-99,共4页
Ship Electronic Engineering
关键词
合成孔径雷达
目标识别
多特征决策融合
synthetic aperture radar
target recognition
multi-feature decision fusion
作者简介
马梓元,男,硕士研究生,研究方向:先进飞行控制技术;龚华军,男,教授,博士生导师,研究方向:先进飞行控制技术、光传飞控技术;王新华,男,副教授,硕士生导师,研究方向:先进飞行控制技术;刘禹,女,硕士研究生,研究方向:先进飞行控制技术。
