摘要
针对合成孔径雷达(synthetic aperture radar,SAR)图像变化检测获得有标记样本的数量十分有限且困难,传统方法检测率低等问题,提出了一种基于原始特征空间的K均值和支持向量机(K-means and support vector machine,KM-SVM)法SAR图像无监督变化检测。首先,不需要任何先验信息的条件下,利用K-means聚类方法获取差异图像的分类阈值;其次,利用阈值,引入偏移量,自动选取伪训练集和无标签集,并用伪训练集定义SVM的初始决策超平面;最后,用基于统计特征的半监督学习算法和支持向量机相结合对图像进行变化类与非变化类的分类。实验结果表明:该算法优于基于混合高斯分布模型的KI法和基于广义高斯分布模型的KI法,能保持较好的分类、泛化能力和较稳定的检测精度。这些结果表明了文中方法的有效性。
It is difficult to obtain training data in practical synthetic aperture radar （SAR）image change detection tasks, and the detection rates of classical methods are low. A K-means and support vector machine （KMSVM） ap- proach, which aims at extracting the change information in the original feature space without any training data, is pro- posed. First, the threshold of difference images is selected by the K-means clustering method. Second, an offset should be obtained for deriving an unlabeled set and a pseudo-training set that is necessary for initializing a bina- ry support vector machine （SVM） classifier. Finally, according to a semi-supervised learning algorithm based on statistical characteristics, the SVM performs change detection by considering unlabeled data in definition of the decision boundary between unchanged and changed pixels. Experiment results show that, the proposed ap proach achieves better performance, higher generalization ability and more stable change detection precision, than the classical GM-KI, GGM-KI methods. These results prove the efficiency of the proposed approach.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2015年第5期1042-1046,共5页
Systems Engineering and Electronics
基金
国家自然科学基金(61072141
61132008)资助课题
关键词
合成孔径雷达图像
变化检测
半监督学习
K-均值聚类
支持向量机
synthetic aperture radar （SAR）images
change detection
semi-supervised learning
K-means
support vector machine （SVM）
