王海鹏,,
徐丰
复旦大学电磁波信息科学教育部重点实验室 上海 200433
基金项目:国家自然科学基金(61991422)
详细信息
作者简介:郭倩：郭 倩(1996–)，女，山西平遥人，复旦大学电磁波信息科学教育部重点实验室博士研究生，主要研究方向为雷达成像与智能感知技术。E-mail: 18210720055@fudan.edu.cn
王海鹏(1979–)，男，河南遂平人，复旦大学电磁波信息科学教育部重点实验室教授，研究方向为雷达系统设计与算法开发、遥感图像处理与信息获取、机器学习与目标识别、智能图像处理等。E-mail: hpwang@fudan.edu.cn
徐丰：徐 丰(1982–)，男，浙江东阳人，复旦大学博士学位，教授，复旦大学电磁波信息科学教育部重点实验室副主任，研究方向为SAR图像解译、电磁散射建模、人工智能，兼职：IEEE地球科学与遥感快报副主编、IEEE地球科学与遥感学会上海分会主席。E-mail: fengxu@fudan.edu.cn
通讯作者:王海鹏 hpwang@fudan.edu.cn
责任主编：高鑫 Corresponding Editor: GAO Xin中图分类号:TN957.51
计量
文章访问数:2483
HTML全文浏览量:785
PDF下载量:704
被引次数:0
出版历程
收稿日期:2020-03-17
修回日期:2020-05-29
网络出版日期:2020-06-18
Research Progress on Aircraft Detection and Recognition in SAR Imagery
GUO Qian,WANG Haipeng,,
XU Feng
Key Laboratory for Information Science of Electromagnetic Waves, Fudan University, Shanghai 200433, China
Funds:The National Natural Science Foundation of China (61991422)
More Information
Corresponding author:WANG Haipeng, hpwang@fudan.edu.cn
摘要
摘要:目标检测与识别是高分辨合成孔径雷达(SAR)领域的热点问题。机场上飞机作为一种典型目标，其检测和识别有一定的独特性。该文回顾了SAR图像典型目标检测识别领域技术的发展过程，分析了SAR图像中飞机目标的散射机制及面临的技术难点，阐述了 SAR 飞机目标检测识别的系统流程、技术路线和关键科学问题，对基于传统与基于深度学习两个方面的飞机目标检测识别的研究进展进行了归纳总结，并讨论了各类方法的特点及存在的问题，展望了未来的发展趋势。该文认为如何将深度学习与目标电磁散射机理结合、提高网络或模型的泛化能力是提升SAR图像中目标检测识别精度的关键，并给出了一种基于散射信息与深度学习融合的飞机目标检测方法。
关键词:合成孔径雷达/
飞机检测/
飞机识别/
散射信息/
深度学习
Abstract:Target detection and recognition are popular issues in the field of high-resolution Synthetic Aperture Radar (SAR). As a typical target, aircraft detection and identification has certain uniqueness. This paper reviews the development of detection and recognition techniques for a typical target in SAR imagery, analyzes the scattering mechanism and technical difficulties of aircraft in SAR imagery, describes the system flow, technical routes, and key scientific problems of target aircraft detection and recognition in SAR imagery, summarizes the research progress from traditional methods to deep-learning-based methods for aircraft detection and recognition, discusses the characteristics and existing problems of various methods, and predicts the future development trend. This paper proposes that combining target electromagnetic scattering mechanism with deep convolutional neural network to improve the generalization capability of the model is the key to improve SAR detection and recognition performance. Moreover, this paper establishes an aircraft detection method based on the fusion of scattering information and deep convolutional neural network.
Key words:Synthetic Aperture Radar(SAR)/
Aircraft detection/
Aircraft recognition/
Scattering information/
Deep learning
PDF全文下载地址:
https://plugin.sowise.cn/viewpdf/198_236fc0b3-cc7f-436d-84b6-19afd7bdf304_R20020
