This paper systematically studies the convergence behavior of rank filters. The problem of convergence behavior of rank filters has been solved completely for bounded sequences. Moreover, some properties of its limiti...This paper systematically studies the convergence behavior of rank filters. The problem of convergence behavior of rank filters has been solved completely for bounded sequences. Moreover, some properties of its limiting sequences and recurrent sequences are obtained.展开更多
为了避免图像辐射失真带来的对应点匹配失效,达到更好地标准化双目摄像机的平行结构,文中提出一种融合改进Rank变换和LoG(Laplacian of Gauss)滤波的RL滤波用于左右视图像素域的亮度变换,并利用极线校正实现图像的几何变换;经过预处理后...为了避免图像辐射失真带来的对应点匹配失效,达到更好地标准化双目摄像机的平行结构,文中提出一种融合改进Rank变换和LoG(Laplacian of Gauss)滤波的RL滤波用于左右视图像素域的亮度变换,并利用极线校正实现图像的几何变换;经过预处理后,边缘信息、亮度大小的统计信息并结合高斯平滑滤波器来降低辐射失真和噪声的影响,并能够将匹配点的搜索范围降到1D,提高了后序匹配的可执行性和鲁棒性,相对于LoG滤波和Rank变换,总体匹配精度提高了14.74%和22.36%。展开更多
基金This work was supported by the China Doctoral Research Foundation and the National Natural Science Foundation of China(Grant No.10171050).
文摘This paper systematically studies the convergence behavior of rank filters. The problem of convergence behavior of rank filters has been solved completely for bounded sequences. Moreover, some properties of its limiting sequences and recurrent sequences are obtained.
文摘为了避免图像辐射失真带来的对应点匹配失效,达到更好地标准化双目摄像机的平行结构,文中提出一种融合改进Rank变换和LoG(Laplacian of Gauss)滤波的RL滤波用于左右视图像素域的亮度变换,并利用极线校正实现图像的几何变换;经过预处理后,边缘信息、亮度大小的统计信息并结合高斯平滑滤波器来降低辐射失真和噪声的影响,并能够将匹配点的搜索范围降到1D,提高了后序匹配的可执行性和鲁棒性,相对于LoG滤波和Rank变换,总体匹配精度提高了14.74%和22.36%。
