Thresholding value getting out of limit

Hey guys I am new to image processing and I want to convert my image to binary using some threshold value which consists of image mean and standard deviation. Now problem is this threshold value gets more than 255 (I am using Grey scale image) so there is nothing remain in image for binary conversion only black image comes so can anybody help how to do it.I am writing my code below::

%%%%%%%code starts %%%%%%%

(1)I=imread('testimage')

(2)mean_image=mean2(I)

(3)std_image=std2(I).....

%%%%%%%%%%%%%%%code ends :::::

now i count thresholding

value which goes more than 255 so any idea what to do?

thanks in advance

Answer 1

[Image Analyst]

Please explain why you'd even WANT a threshold that is more than 255. It just doesn't make sense for an image that has values only in the 0-255 range.

[john]

because i after this, i am trying to extract object from it and if thresholding value is more than 255 and I use high pass so only black screen comes.No object can be seen.Tell me if you need more info.

[Image Analyst]

You will never find anything. If your image does not have any pixels with values more than 255, then doing

binaryImage = grayImage > 567; % or whatever

will never, ever, ever, EVER find anything. Sure you can do it but I don't know why since, like I said and can't stress enough, it will never find anything.

[john]

I know there will be no pixel in the image more than 255 but my thresholding value comes like that and because of that only I asked here to know is there any other way?Second thing I have same group of other images which works fine with this thresholding function and even I can extract object also.So is there any way...

[Image Analyst]

You can use a threshold less than 255 to find things:

 binaryImage = grayImage > 42; % or whatever value < 255 you want

If that doesn't answer it, then you'll have to explain what "now i count thresholding" means, plus show your code. Your steps above are not really code. There is nowhere in your "code" that says how you determine a threshold value. Did you do something like

thresholdValue = mean_image + 3 * std_image;

or something similar? I don't know because you didn't share that. Even if you did, it's possible that that formula could give a value more than 255, in which case you might want to investigate whether that is a good formula to use.

[john]

Actually your guess is right I am using mean_image+2*std_image and as I said above my thresholding value only comes above 255 not my mean or standard deviation.Second thing is this thresholding function was mentioned in one of the IEEE paper I am following so that I think author must have used it.So if he doesn't have any problem to find threshold then what I could do to overcome this problem that's what I wanted to know.As I mentioned this formula works for other images but for some images only it gives problem.Tell me if you want to know more.

[Image Analyst]

That image is too bright for that method of finding the threshold. You need to find another method. If you post your image soon, I could take a look at it, but I'll be gone over the weekend.

[john]

Ya your are right.My image is bright but i will appreciate very much if you give me any other method to extract that type of object in image.I am attaching image to this message http://imageshack.us/photo/my-images/9/testimagey.jpg/

[Image Analyst]

You don't really expect to find an object in all that binary noise do you? I see no object whatsoever.