• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>International conference on graphic and image processing >Change Detection for Synthetic Aperture Radar Images Based on Pattern and Intensity Distinctiveness Analysis
【24h】

Change Detection for Synthetic Aperture Radar Images Based on Pattern and Intensity Distinctiveness Analysis

机译：基于模式和强度区分分析的合成孔径雷达图像变化检测

获取原文
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Synthetic aperture radar (SAR) image is independent on atmospheric conditions, and it is the ideal image source for change detection. Existing methods directly analysis all the regions in the speckle noise contaminated difference image. The performance of these methods is easily affected by small noisy regions. In this paper, we proposed a novel change detection framework for saliency-guided change detection based on pattern and intensity distinctiveness analysis. The saliency analysis step can remove small noisy regions, and therefore makes the proposed method more robust to the speckle noise. In the proposed method, the log-ratio operator is first utilized to obtain a difference image (DI). Then, the saliency detection method based on pattern and intensity distinctiveness analysis is utilized to obtain the changed region candidates. Finally, principal component analysis and k-means clustering are employed to analysis pixels in the changed region candidates. Thus, the final change map can be obtained by classifying these pixels into changed or unchanged class. The experiment results on two real SAR images datasets have demonstrated the effectiveness of the proposed method.
机译：合成孔径雷达（SAR）图像不受大气条件的影响，是用于变化检测的理想图像源。现有方法直接分析斑点噪声污染的差分图像中的所有区域。这些方法的性能容易受到较小噪声区域的影响。在本文中，我们提出了一种新的基于模式和强度差异分析的显着性变化检测框架。显着性分析步骤可以去除较小的噪点区域，因此使所提出的方法对斑点噪声更加鲁棒。在提出的方法中，首先使用对数比率运算符来获取差异图像（DI）。然后，利用基于模式和强度差异性分析的显着性检测方法来获得变化区域候选。最后，采用主成分分析和k均值聚类来分析变化区域候选中的像素。因此，可以通过将这些像素分类为改变或未改变的类别来获得最终改变图。在两个真实SAR图像数据集上的实验结果证明了该方法的有效性。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号