boxfilter 实现

A implementation of boxfilter

boxfilter 是均值滤波的一种改进。在以下这篇blog里面有介绍。

http://www.cnblogs.com/easymind223/archive/2012/11/13/2768680.html

这里我使用matlab对其进行实现。

可是这里边界问题没有处理好，假设viewer有好的意见欢迎告诉我

% % *********************************************************
% code writer      : EOF
% code file        : my_boxfilter.m
% code date        : 2014.10.21
% e-mail           : jasonleaster@gmail.com
%
% Code description :
%           Here is my implementation of boxfilter :)
%  It work correctly in the region where is not close to
%  the end boundary but will meet problem on some where
%  close to the image end boundary.
% *********************************************************
 
function Filted_Img = my_boxfilter(Image)
 
    if size(Image,3) ~= 1
        fprintf('ERROR Imput-Image must be ##ONE## channel image\n');
        return;
    end
 
    Height_Img = size(Image,1);
    Width_Img  = size(Image,2);
 
    Buffer     = zeros(1,Width_Img);
    Filted_Img = zeros(Height_Img,Width_Img);
 
    % treat this varible as a constant
    SEARCH_WIN_HEIGHT = 10;
    SEARCH_WIN_WIDTH  = 10;
 
    for row = 1: Height_Img
        for col  = 1: Width_Img
 
            sum_value = 0;
            if (row + SEARCH_WIN_HEIGHT) < Height_Img
 
                for temp = row : (row + SEARCH_WIN_HEIGHT)
                    sum_value = sum_value + Image(temp,col);
                end
            else
 
                for temp = row : Height_Img
                    sum_value = sum_value + Image(temp,col);
                end
            end
 
            Buffer(col) = sum_value;
        end
 
        for col = 1:Width_Img
 
            if (col + SEARCH_WIN_WIDTH) < Width_Img
 
                for temp = col : col + SEARCH_WIN_WIDTH
                        Filted_Img(row,col) =  Filted_Img(row,col) + Buffer(temp);
                end
            else
                for temp = col : Width_Img
                        Filted_Img(row,col) =  Filted_Img(row,col) + Buffer(temp);
                end
            end
        end
 
    end
 
    Filted_Img = Filted_Img./(SEARCH_WIN_HEIGHT * SEARCH_WIN_WIDTH);
end

版权声明：本文博客原创文章，博客，未经同意，不得转载。