摘要
针对基于差异图像分类的SAR图像变化检测方法中差异图像保持完整变化区域难的特点,以及按像元分类容易受噪声的干扰,提出了一种基于综合差异图像和按块k均值聚类法的SAR图像变化检测方法。首先,分别通过差值法和对数比值法得到2幅不同时相同一地理位置的SAR图像差异图像,为了使差异图像更加平滑和保持边缘信息,分别对这2种差异图像进行均值滤波和中值滤波。然后,通过简单的线性结合得到最终的差异图像,随后将差异图像分成若干个大小为h×h且不重叠的块,通过主成分分析提取每个块的特征向量,再利用k均值聚类法将特征向量空间分成2类。最后,根据最近邻法将差异图像分为变化区域和未变化区域。实验结果表明,该方法不仅能有效地检测出变化区域,还在一定程度上降低了虚警。
At present, the approaches of change detection in SAR images based on difference image classification are prevalent, but it is usually difficult to preserve complete changed area in difference images and the method of classification based on pixel is easy to be disturbed by noises. In this paper, we propose a simple yet effective change detection in synthetic aperture radar images using combined difference image and block-based k-means clustering. Firstly, the subtraction operator and the log ratio are applied to generate two kinds of simple difference images. Secondly, the mean filter and the median filter are used to the two difference images, respectively, where the means filter focuses on making the difference image smooth, and the median filter is used to preserve the edge information. And then, a simple combination framework which uses the images obtained by the mean filter and the median filter is proposed to generate a better difference image. Next, the difference image is partitioned into h × h non-overlapping blocks, and orthonormal eigenvectors are extracted through principal component analysis of h × h non-overlapping block set to create an eigenvector space. Finally, the change detection is achieved by partitioning the feature vector space into two clusters using k-means clustering with k = 2 and then each pixel is assigned to one of the two clusters by using nearest neighbor method. Experimental results confirm the effectiveness of the proposed approach, and false alarm is reduced.
作者
丁翔
沈汀
DING Xiang SHEN Ting(Institute of Remote Sensing and Digital Earth, Chinese Academy of Sciences, Beijing 100094 University of Chinese Academy of Sciences ,Beijing 100049)
出处
《遥感信息》
CSCD
北大核心
2017年第4期47-51,共5页
Remote Sensing Information
关键词
变化检测
差异图像
主成分分析
K均值聚类
SAR图像
change detection
difference image
principal component analysis
k-means clustering
synthetic aperture radar image
