linsolve 中的错误

Question

A=imread('lena_color.jpg');
x1=2.1;
b=A*x1;
b=double(b);
x1=double(x1);
opts.UT = true; opts.TRANSA = false;
A1 = linsolve(x1,b,opts);
figure;
imshow(A1);

这给出了：

 error >>??? Error using ==> linsolve
First and second arguments must be single or double.

Error in ==> test at 9
A1= linsolve(b,x1,opts);

请帮助解决这个问题。

还有其他方法可以求解Eq1= A*2.1 + B*3.5 +C*1.5 形式的方程吗？

Answer 1

我认为问题在于尺寸，而不是论据。 以下代码适用于我：

% you can switch it with your own image after you see it works
I = imread('cameraman.tif');
A = im2double(I);

% scalar*matrix works in matlab commandline,
% but needs to be defined when it comes to equations
x1=2.1*eye(size(A));
b=A*x1;

opts.UT = true; 
opts.TRANSA = false;
A1 = linsolve(x1,b,opts);

figure; imshow(A1);

[编辑]@SKM 和@gary，这里有一个详细的解决方案（对于 RGB 情况）：

% example for RGB image
Img= im2double( imread('peppers.png') );
Dim = size(Img, 3);

% transform the original image, 
% by multiplying every color channel by different scalar.
% that is going to be a very red image...
a = [1 0.5 0.2]; 
aMat = []; newImg = [];
for Ind=1:Dim
    aMat(:, :, Ind) = a(Ind)*eye( size(Img, 1), size(Img, 2) );
    newImg(:, :, Ind) = a(Ind)*squeeze(Img(:, :, Ind));
end

% recontructing original image
recontrcutedImg=[];
for Ind=1:Dim
    recontrcutedImg(:, :, Ind) = linsolve( aMat(:, :, Ind), newImg(:, :, Ind) );
end

% show the images
figure;
subplot(1, 3, 1); imagesc(Img); title('original');
subplot(1, 3, 2); imagesc(newImg); title('changed image');
subplot(1, 3, 3); imagesc(recontrcutedImg); title('reconstructed image');