摘要
合成孔径雷达(synthetic aperture radar, SAR)图像分类是遥感领域最重要的课题之一。然而,SAR图像特征提取的困难和相干斑噪声的存在都严重影响了SAR图像分类的准确性。为了克服这些问题,文章提出了一种新的SAR图像分类算法。该算法将相干斑去噪技术和深度置信网络相结合,在通过深度置信网络对SAR图像进行无监督的学习和特征提取的同时,提出了区域滤波的方法来减少相干斑噪声对分类结果的影响。实验采用了不同噪声水平的合成SAR图像以及由RADARSAT-2获取的真实SAR图像进行测试。实验结果表明,与传统的分类方法相比,该算法在噪声鲁棒性和分类能力方面都有良好的改进;同时,该算法在边界区域具有优秀的分类能力。
Synthetic aperture radar(SAR) image classification is one of the most important topics in remote sensing. However, the absence of discriminative features and the existence of speckle noise in SAR images severely affect the classification of SAR images. In order to overcome these problems, a novel SAR image classification algorithm combining deep belief network and region filter is proposed. The region filter is proposed to reduce speckle noise while preserving boundary information. Then, through the unsupervised feature learning and supervised fine-tuning of deep belief network, the classification is well completed. To verify the effectiveness of the proposed method, it is tested on both simulated SAR and real RADARSAT-2 SAR images with varying noise levels. Experimental results demonstrate that compared with the traditional methods, the proposed method has fine improvements in noise immunity and classification accuracy. Moreover, the proposed algorithm shows better classification ability in boundary areas.
作者
夏晶凡
杨学志
贾璐
XIA Jingfan;YANG Xuezhi;JIA Lu(Intelligent Manufacturing Institute,Hefei University of Technology,Hefei 230051,China;School of Computer and Information,Hefei University of Technology,Hefei 230601,China)
出处
《合肥工业大学学报（自然科学版）》
CAS
北大核心
2019年第12期1636-1643,共8页
Journal of Hefei University of Technology：Natural Science
基金
国家自然科学基金资助项目(61371154
61701157
41601452)
安徽省重点研究与开发计划资助项目(1704a0802124)
安徽省自然科学基金资助项目(1608085QF142)
关键词
SAR图像分类
深度置信网络
区域识别
边界保持
抗噪
synthetic aperture radar(SAR)image classification
deep belief network
region recognition
boundary preservation
noise immunity
