摘要
合成孔径雷达具有全天时、全天候、远作用距离、高分辨等特点,在各个领域发挥了重要作用。基于SAR图像的分类识别技术,特别是地面相关目标识别技术受到广泛关注。针对单一的特征很难完全描述SAR图像中感兴趣目标,提出了一种基于主成分分析(PCA)降维的特征融合分类方法。首先采用多种特征尽可能丰富地描述目标;其次用PCA方法对冗杂特征进行降维,得到降维后的特征;最后将其放入支持向量机进行分类,本文使用MSTAR公开数据集进行了数值实验,遍历参数寻优并输出最终的分类结果。在预处理阶段,本文研究了均值滤波降噪方法对SAR图像乘性噪声的影响,通过实验验证是否需要将SAR图像转化为加性噪声再进行滤波处理。最终实验结果显示,提出的方法能够达到较高的分类正确率,证明了该方法的有效性,且相比较将乘性噪声转化为加性噪声,直接使用均值滤波效果更好。
Synthetic aperture radar(SAR)has the characteristics such as all-day,all-weather,long range,and high resolution,and plays an important role in various fields.The classification and recognition technology based on SAR images,especially the ground related target recognition technology,has attracted extensive attention.Since it is difficult for a single feature to fully describe the target of interest in an SAR image,a feature fusion classification method based on principal component analysis(PCA)dimensionality reduction is proposed.First,multiple features are used to describe the target as richly as possible.Second,the PCA method is used to reduce the dimensionality of redundant features to obtain the features after dimensionality reduction.Finally,the obtained futures are put into the support vector machine(SVM)for classification.In this paper,the MSTAR public data set is used to carry out numerical experiments,traverse the parameter optimization,and output the final classification results.In addition,in the preprocessing stage,the effects of the mean filtering noise reduction method on the multiplicative noise of SAR images are studied,and the necessity of transforming SAR images into additive noise for filtering is verified.The final experimental results show that the method proposed in this paper can achieve high classification accuracy.This proves the effectiveness of this method.It shows that compared with converting multiplicative noise into additive noise,the direct use of mean filtering is better.
作者
陶佳慧
别雨轩
顾约翰
王辉
周维
TAO Jiahui;BIE Yuxuan;GU Yuehan;WANG Hui;ZHOU Wei(Shanghai Institute of Satellite Engineering,Shanghai 201109,China;Shanghai Institute of Spaceflight Control Technology,Shanghai 201109,China)
出处
《上海航天（中英文）》
CSCD
2021年第S01期98-102,共5页
Aerospace Shanghai（Chinese&English）
基金
“上海市毫米波空天信息获取及应用技术重点实验室”项目资助
关键词
支持向量机
PCA降维
多特征融合
参数寻优
support vector machine(SVM)
principal component analysis(PCA)dimensionality reduction
multi-feature fusion
mean filtering
作者简介
陶佳慧(1996—),男,硕士,主要研究方向为SAR数据处理与应用。
