Message passing Interface (MPI) is a kind of network distributed parallel computing environments whichhave been widely used on super parallel computers and networks. First,this paper describes the rssearch backgrounda...Message passing Interface (MPI) is a kind of network distributed parallel computing environments whichhave been widely used on super parallel computers and networks. First,this paper describes the rssearch backgroundand developing status of MPI. Then on this basis it will study and analyze the functions and features of MPI ,summa-rize its insufficiencies and gives some suggestions for modification.展开更多
针对可见光图像弱小目标检测中的背景抑制和去噪问题,提出了奇异值分解(Singular Value Decomposition,SVD)带通滤波新方法.首先分析了图像奇异值与目标、噪声和图像背景的关系,结果表明奇异值的高序部分更多地反映图像噪声,中序部分更...针对可见光图像弱小目标检测中的背景抑制和去噪问题,提出了奇异值分解(Singular Value Decomposition,SVD)带通滤波新方法.首先分析了图像奇异值与目标、噪声和图像背景的关系,结果表明奇异值的高序部分更多地反映图像噪声,中序部分更多地反映目标性质,而低序部分更多地反映图像背景.以此为依据提出了SVD-Ⅰ型和SVD-Ⅱ型两种带通滤波器,并给出了奇异值曲线转折点法和门限准则法两种滤波器参数确定方法.实验表明SVD带通滤波能有效抑制图像背景,去除噪声,进而提高弱小目标的信噪比.展开更多
文摘Message passing Interface (MPI) is a kind of network distributed parallel computing environments whichhave been widely used on super parallel computers and networks. First,this paper describes the rssearch backgroundand developing status of MPI. Then on this basis it will study and analyze the functions and features of MPI ,summa-rize its insufficiencies and gives some suggestions for modification.
文摘针对可见光图像弱小目标检测中的背景抑制和去噪问题,提出了奇异值分解(Singular Value Decomposition,SVD)带通滤波新方法.首先分析了图像奇异值与目标、噪声和图像背景的关系,结果表明奇异值的高序部分更多地反映图像噪声,中序部分更多地反映目标性质,而低序部分更多地反映图像背景.以此为依据提出了SVD-Ⅰ型和SVD-Ⅱ型两种带通滤波器,并给出了奇异值曲线转折点法和门限准则法两种滤波器参数确定方法.实验表明SVD带通滤波能有效抑制图像背景,去除噪声,进而提高弱小目标的信噪比.
