摘要
针对大幅面合成孔径雷达(SAR)图像的斑点抑制问题,提出了一种基于自适应模板的梯度倒数加权(GIW)算法,在编程实现上采用图形处理器(GPU)进行优化,有效地解决了由于像素级运算复杂度高所导致的大幅面雷达图像处理实时性差的问题。高分辨率SAR图像的处理结果显示,该算法在有效抑制斑点噪声的同时较好地保持了边缘细节信息。经过GPU加速后,对于大幅面图像的处理,相对于中央处理器(CPU)实现可以有2个量级以上的速度提升。
For the denoising of large SAR（synthetic aperture radar） image,a modified GIW（gradient inverse weighted） algorithm with auto-adaptive template is proposed.And then an optimal implementation based on the GPU（graphic processing unit） is given,which solves the problem of real-time veracity by massive computation of pixel-level calculation.The experimental results with the high resolution SAR images indicate the nice efficiency in the noise restraintation and edge preservation,as well as the prominent computation speed by the GPU,which can be twice level faster than that of by CPU（central processing unit） for large image.
出处
《计算机工程与设计》
CSCD
北大核心
2010年第15期3455-3458,3502,共5页
Computer Engineering and Design
关键词
图形处理器
合成孔径雷达图像
噪声抑制
梯度倒数加权算法
算法优化实现
graphic processing unit
synthetic aperture radar image
noise reduction
gradient inverse weighted algorithm
optimal implementation
