how to compute min for image with black border?

Hi, I have an image with black border in 2000 * 2000.after using min function the output is 0 or 1.that The minimum is related to border of the image that is black.how to compute min or max other than black border? my code is here.by this code output min is NaN.

    img=imread('image.tif');
    figure(1),imshow(img),title('original');
    [M,N]=size(img);
     for i=1:M
     for j=1:N
     if  img(i,j)<1
     img(i,j)=NaN;
    end
    end
    end     
    minofimg=min(min(img));

Answer 1

[Chad Greene]

What about

min(img(img~=0))

That is, the minimum value of img values that are not zero.

[fred bnm]

ok.thanks.Is it possible to access pixels Apart from zero intensity by using binary mask ?

%msk is binary mask
image=img(img~=msk);

[Chad Greene]

Yep, looks good to me.

[Image Analyst]

Looks bad to me.

For one thing, "image" is a reserved built-in function and you never want to use that for a variable name. Secondly, if the min intensity value of your image, not including the black surrounding border, is zero, the above code will not tell you that it's zero. You'd get the next darkest value, which is wrong. Better (more robust) is to find the border and crop it off and then use min().

[Chad Greene]

Ah, very good points!

[fred bnm]

how to find the border and crop it off?

[Chad Greene]

You can specify the range of rows and columns you want to keep. For example start with this image:

img = imread('cameraman.tif'); 
image(img)
colormap(gray)

And clip the first 49 rows, and the last 47 rows. Also clip the first 124 columns:

keeprows = 50:209; 
keepcols = 125:256; 
img_trimmed = img(keeprows,keepcols); 
image(img_trimmed)