How can we superimpose more than 2 images?

For e.g; imshowpair(Image1,image2); can be use to superimpose 2 images. What if I have 10 images?

[Geoff Hayes]

Hina - my understanding of http://www.mathworks.com/help/images/ref/imshowpair.html is that it shows differences between two images which (to me) is different from superimposing two images. Which is it that you want to do?

[Adam]

imshowpair looks like a convenience function to do things that you can easily do yourself also if you wish.

e.g. the 'montage' option can be replicated using as many images as you want simply by doing (to expand on the example in the help doc):

A = imread('cameraman.tif');
B = imrotate(A,5,'bicubic','crop');

myImage = [A B A B A B];

figure; imshow( myImage );

The blending options require more thought though, especially for more than 2 images as defining a blending of 3 or more images can get quite complicated and understanding what you are looking at certainly can.

[Hina Ismail]

@Geoff @Adam thanks for your response.

I1 = imread('Capture1.jpg');

BW1 = edge(I1,'sobel');

%hold on
I2 = imread('Capture3.jpg');

BW2 = edge(I2,'sobel');
%hold on
I3 = imread('Capture5.jpg');

BW3 = edge(I3,'sobel');

title('Sobel Filter') 
% hold all
figure(1), 
imshowpair(BW1,BW2);

Thats my coding for two images, What IF I want to add BW3 in the same figure to see the difference? .

[Hina Ismail]

Please have a look the image. Similarly I want to see edge difference of all images. Hope my question is clear now. Thanks once again

[Adam]

You need to be clear as to what you are aiming to do, independent of the method you will use to do it.

The difference of two images is clear and easy to define. As soon as you add a 3rd image this is no longer the case.

How do you mathematically define the difference of 3 images, let alone 10 images?

Once you have answered that then what you want to do is trivial because you don't need imshowpair, you simply apply the maths yourself and use imshow on the compound result.

Answer 1

[Image Analyst]

Average them

sumImage = double(image1)+ double(image2)+ double(image3)+ double(image4) + ...... etc.
meanImage = sumImage / numberOfImages;

Answer 2

[Hina Ismail]

Hi Guys,

Thanks for your response. The attached figure is the edge detection of two images, where I can see the difference of edges. Similarly, I have 10 images how can I show it in the same image?

[Image Analyst]

Was there something about summing and dividing by the number of images that you didn't understand?

Anyway, I don't see how your image is the " the edge detection of two images". Why don't you show the two gray scale images, and the two edge images. All I see is a single annotated image.

[Hina Ismail]

Thanks for response. Summing of images does work. Is there anyway I can color each edges to indicate different? for e.g; image edge 1 should be green color, image edge 2 red color and so on.

[Hina Ismail]

Another question. Is there anyway I can show the calculated edge difference of each image? May be by calculating difference pixels?

[Image Analyst]

It looked like your edges are in different places, so exactly what does the "edge difference" mean? You can just subtract the edge images and that will make sense if the edges are in the same place but just have different strengths, but if they are in different locations, then what does that mean?

If you want to threshold the images and assign each binary image to a different value, then you can simply add them together and apply a colormap with label2rgb().