Question

我有一个带有6个波段的多光谱图像。

imagen1=imread('re_2008.tif')
size2=size(imagen1);
nrow= size2(1);
ncol= size2(2);
nband= size2(3);

我想要做的是将每个波段的每个像素（在相同位置）与其值相关联，进行插值并使用位于其他波长的新值替换它。如果我向你展示代码，也许你更了解我。

imagen3_2 = zeros(nrow, ncol, nband);
var1= [1 2 3 4 5 6]'; %'

for row=1:nrow;
    for column=1:ncol;
        for band=1:nband;              
            v = imagen1(nrow(1),nband(2),:); v = v(:); 
            t= 0:100;

            interplan= interp1(var1, v, t,'cubic');
            y5 = interplan(5); % I get the value of the interpolation on this position
            y12 = interplan(12);
            y20 = interplan(20);
            y50 = interplan(50);
            y80 = interplan(80);
            y90 = interplan(90);

            imagen3_2(:,:,1)= (y5); % and I replace it
            imagen3_2(:,:,2)= (y12);
            imagen3_2(:,:,3)= (y20);
            imagen3_2(:,:,4)= (y50);
            imagen3_2(:,:,5)= (y80);
            imagen3_2(:,:,6)= (y90);
        end 
    end
end

我得到相同的值，而不是每个像素。

提前致谢，

Answer 1

怎么样

imagen3_2 = zeros(nrow, ncol, nband);
var1= [1 2 3 4 5 6]'; %'

for row=1:nrow;
    for column=1:ncol;
            v = imagen1(row, column, :); v = v(:); 
            t= 0:100;

            interplan= interp1(var1, v, t,'cubic');
            y5 = interplan(5); % I get the value of the interpolation on this position
            y12 = interplan(12);
            y20 = interplan(20);
            y50 = interplan(50);
            y80 = interplan(80);
            y90 = interplan(90);

            imagen3_2(row,column,1)= (y5); % and I replace it
            imagen3_2(row,column,2)= (y12);
            imagen3_2(row,column,3)= (y20);
            imagen3_2(row,column,4)= (y50);
            imagen3_2(row,column,5)= (y80);
            imagen3_2(row,column,6)= (y90);
    end
end

Answer 2

我看到Shai已经回答了你的问题，但如果你对变量y5，y12等并不感兴趣，那么有一种方法可以让代码更紧凑，更容易维护：

imagen3_2 = zeros(nrow, ncol, nband);
var1= [1 2 3 4 5 6]'; %'

for row=1:nrow;
    for column=1:ncol;
            v = imagen1(row, column, :); v = v(:); 
            t= 0:100;

            interplan= interp1(var1, v, t,'cubic');

            y = [5 12 20 50 80 90];
            for i = 1:length(y)         
               imagen3_2(row,column, i)= interplan(y(i));% repace interpolation value
            end
    end
end

Answer 3

为了提高效率，尽量避免使用Matlab中的循环，你可以试试这样的东西：

x = [1 2 3 4 5 6];
y = imread('re_2008.tif');
y_size = size((y(:,:,1)));

xi = 0:100;
yi = zeros([y_size,numel(xi)]);

for i_pix=1:prod(y_size)
    [aux_x,aux_y] = ind2sub(y_size,i_pix);
    yi(aux_x,aux_y,:) = interp1(x,squeeze(y(aux_x,aux_y,:)),xi,'cubic');
end

或者甚至喜欢这样：

x = [1 2 3 4 5 6];
y = imread('re_2008.tif');
xi = 0:100;

yi = arrayfun(@(y1,y2,y3,y4,y5,y6) interp1(x,[y1,y2,y3,y4,y5,y6],xi,'cubic'), ...
y(:,:,1), y(:,:,2), y(:,:,3), y(:,:,4), y(:,:,5), y(:,:,6))

我无法检查代码，因此可能存在错误;但它只是提供有关其他类型实施的想法。

多光谱图像的Píxel环

3 个答案: