基于Rsnet框架和Mobilenet迁移网络的遥感机场目标识别  

Remote Sensing Airport Target Recognition Based on Rsnet Framework and Mobilenet Migration Network

作　　者：李昂[1,2] 王晟全[2] 张晨 LI Ang;WANG Shengquan;ZHANG Chen(College of Telecommunications&Information Engineering,Nanjing University of Posts and Telecommunications,Nanjing 210003,China;Institute of Electronic Engineering and Optoelectronic Technology,Nanjing University of Science and Technology ZiJin College,Nanjing 210003,China)

机构地区：[1]南京邮电大学通信学院,南京210003 [2]南京理工大学紫金学院,南京210003

出　　处：《重庆理工大学学报（自然科学）》2021年第6期131-136,298,共7页Journal of Chongqing University of Technology：Natural Science

基　　金：江苏省高校自然科学基金面上项目(18KJD510004);江苏省普通高校学术学位研究生科研创新计划项目(KYLX16_0661);江苏省高等学校大学生创新创业训练计划项目(201913654002Y)。

摘　　要：为精准高效地识别机场目标遥感图像,提出基于Rsnet框架和Mobilenet迁移网络的方法,在特征不明显或目标特征像素点较低的情况下,提高了Map值。结果表明:在同等条件下,不但能够取得较高的Map值,还能够提高采样率,节约了训练时间。对于实际目标识别具有理论意义和实用价值。In order to identify the remote sensing image of airport target accurately and efficiently,this paper proposes a method based on Rsnet framework and Mobilenet migration network,which mainly solves the problem that the Map value is not high when the features are not obvious or the target feature pixels are low.The results show that under the same conditions,not only can we get a higher Map value,but also can improve the sampling rate and save the training time.Therefore,it has a certain theoretical significance and practical value for the practical application of target recognition.

关 键 词：Rsnet mobilenet 目标检测 深度学习 机场目标 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]