摘要
提出了一种支持向量机和隐马尔可夫模型相结合的合成孔径雷达图像目标识别方法。该方法用小波分解和主成分分析提取图像特征,生成特征向量。利用图像在方位角上的关系由特征向量生成图像的特征序列以及隐马尔可夫模型的训练序列。用支持向量机进行目标预识别,确定目标最有可能所属的两个类别,用隐马尔可夫模型在这两个类别中确定目标最终所属类别,完成目标识别。使用MSTAR数据库中的图像数据对该方法进行验证和分析,结果表明,该方法可以明显提高目标的正确识别率,是一种有效的合成孔径雷达图像目标识别方法。
A method for synthetic aperture radar images target recognition using support vector machine combined with hidden Markov models is presented. Feature vectors are acquired by wavelet decomposition and principal component analysis. Feature sequences and training sequences for hidden Markov models are generated from feature vectors using the relationship of multi-images in aspect value. The support vector machine is used to perform pre-recognition in order to decide the two most probable classes the target may belong to. The hidden Markov models are used to determine which class of the two the target belongs to at last. Image samples of
targets in MSTAR database are used to verify the method, and the results show that the proposed method can enhance the target recognition rate evidently and it is an effective method for synthetic aperture radar images target recognition.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2008年第3期447-451,共5页
Systems Engineering and Electronics
关键词
合成孔径雷达
识别
支持向量机
隐马尔可夫模型
synthetic aperture radar
recognition
support vector machine
hidden Markov model