针对双边滤波在抑制SAR图像相干斑噪声的不足,本文提出了一种基于背景匀质性的改进双边滤波算法BH-IBF(Improved Bilateral Filtering algorithm based on Background Homogeneity),并将其应用于SAR图像斑点噪声抑制。BH-IBF以传统双边...针对双边滤波在抑制SAR图像相干斑噪声的不足,本文提出了一种基于背景匀质性的改进双边滤波算法BH-IBF(Improved Bilateral Filtering algorithm based on Background Homogeneity),并将其应用于SAR图像斑点噪声抑制。BH-IBF以传统双边滤波作为基本框架,并利用了双边滤波器中的双边核函数描述像素灰度值之间的相似性以及相邻像素间的几何空间信息。然而,传统双边滤波存在不能有效地滤除强斑点噪声的缺点,并且SAR图像又因成像原理的缺陷导致强斑点噪声普遍存在。针对这些问题,BH-IBF设计了一种根据背景窗口的匀质性进行自适应样本截断的方法,并根据描述背景匀质性的指标自动获取样本截断的截断深度。此外,本文将自适应滤波窗口尺寸以及权重核修改的方案应用到BH-IBF中,以增强匀质区域的斑点噪声平滑强度以及异质区域的边缘信息效果。最后,使用自适应截断后的样本作为已调整权重核的双边滤波器的输入。实验数据显示,BH-IBF能够在有效保留SAR图像纹理信息的同时,获取较好的斑点噪声平滑性能。展开更多
Image registration is the overlaying of two images of the same scene taken at different times or by different sensors. It is one of the essential steps in information processing in remote sensing. To attain a highly a...Image registration is the overlaying of two images of the same scene taken at different times or by different sensors. It is one of the essential steps in information processing in remote sensing. To attain a highly accurate, reliable and low computation cost in image registration a suitable and similarity metric and reduction in search data and search space is required. In this paper, the author shows that if the right bin size is chosen, mutual information can be more robust than correlation in the registration of multi-temporal images. The author also compares the sensitivity of mutual information and correlation to Gaussian and multiplicative speckle noise. The author investigates automatic subimage selection as a reduction in search data strategy. The author proposes a measure, called alienability, which shows the ability ofa subimage to provide reliable registration. Alternate subimage selection methods such as using gradient, entropy and variance are also investigated. The author furthermore looks into a search space strategy using a gradient approach to maximize mutual information and show our first results.展开更多
文摘针对双边滤波在抑制SAR图像相干斑噪声的不足,本文提出了一种基于背景匀质性的改进双边滤波算法BH-IBF(Improved Bilateral Filtering algorithm based on Background Homogeneity),并将其应用于SAR图像斑点噪声抑制。BH-IBF以传统双边滤波作为基本框架,并利用了双边滤波器中的双边核函数描述像素灰度值之间的相似性以及相邻像素间的几何空间信息。然而,传统双边滤波存在不能有效地滤除强斑点噪声的缺点,并且SAR图像又因成像原理的缺陷导致强斑点噪声普遍存在。针对这些问题,BH-IBF设计了一种根据背景窗口的匀质性进行自适应样本截断的方法,并根据描述背景匀质性的指标自动获取样本截断的截断深度。此外,本文将自适应滤波窗口尺寸以及权重核修改的方案应用到BH-IBF中,以增强匀质区域的斑点噪声平滑强度以及异质区域的边缘信息效果。最后,使用自适应截断后的样本作为已调整权重核的双边滤波器的输入。实验数据显示,BH-IBF能够在有效保留SAR图像纹理信息的同时,获取较好的斑点噪声平滑性能。
文摘Image registration is the overlaying of two images of the same scene taken at different times or by different sensors. It is one of the essential steps in information processing in remote sensing. To attain a highly accurate, reliable and low computation cost in image registration a suitable and similarity metric and reduction in search data and search space is required. In this paper, the author shows that if the right bin size is chosen, mutual information can be more robust than correlation in the registration of multi-temporal images. The author also compares the sensitivity of mutual information and correlation to Gaussian and multiplicative speckle noise. The author investigates automatic subimage selection as a reduction in search data strategy. The author proposes a measure, called alienability, which shows the ability ofa subimage to provide reliable registration. Alternate subimage selection methods such as using gradient, entropy and variance are also investigated. The author furthermore looks into a search space strategy using a gradient approach to maximize mutual information and show our first results.
