How to plot intensity profile of an image

I want to plot the intensity profile of this image. I tried using improfile but I don't really know how it works. Can anyone help?

[Walter Roberson]

An intensity profile along which direction?

If you want a profile along a horizontal or vertical line then just use array indexing to extract the vector, and then plot it.

Answer 1

[Akira Agata]

Like this?

I = imread('Capture.PNG');

x = [0 size(I,2)];
y = [size(I,1)/2 size(I,1)/2];
c = improfile(I,x,y);

figure
subplot(2,1,1)
imshow(I)
hold on
plot(x,y,'r')
subplot(2,1,2)
plot(c(:,1,1),'r')
hold on
plot(c(:,1,2),'g')
plot(c(:,1,3),'b')

[Viron Gil Estrada]

Yes! Exactly like that. Would it also work if I change the image to grayscale? Thanks

[Akira Agata]

Thanks, I'm glad to hear that!

If you use grayscale image, improfile function returns N-by-1 vector. So you should change the last 4 lines by plot(c).

[Viron Gil Estrada]

Thanks a lot! That's all I need.

[Warid Islam]

Hi,

How can we plot the intensity profile of the same image along the vertical direction instead of the horizontal direction? Thank you.

[Elif]

How can we measure the areas under these curves?

Answer 2

[Akira Agata]

Hi Warid-san,

You can do it by the same way, like the following.

I = imread('Capture.PNG');

% Since size(I,2)/2 = 215.5, I have set the value '215' directly for x
x = [215 215];
y = [0 size(I,1)];
c = improfile(I,x,y);

figure
subplot(2,1,1)
imshow(I);
hold on
plot(x,y,'r')
subplot(2,1,2)
plot(c(:,1,1),'r')
hold on
plot(c(:,1,2),'g')
plot(c(:,1,3),'b')

[Warid Islam]

Hi Akira,

Thank you very much for the help.

[Manoj Kumar]

Hi,

if there are more than one spot, say 5 (randomly distributed), then how to plot the intensity across all of them?

Thanks

[Akira Agata]

Hi Manoj-san,

Could you upload a sample image file for trying?

[Vivek]

Hi Akira San,How Can you plot the 1D profile of a rectangular region instead of just line ?Tx,vivek

Answer 3

[Lorenzo Gentili]

Hi Akira,

I'm trying to obtain an intesity profile of the image below, but I need more than a line. I would like to have at least 5 vertical line in order to understand how the intesity vary in different point of the image ( lets say the 2 boundaries, the center and 2 point between the boundaries and the center). Is it possible to obtain an unique intensity profile that is the sum of 5 intesity profile obtained by the verical lines ?

Thank you for your help

Lorenzo

[Image Analyst]

To get a mean vertical profile between two columns:

verticalProfile = mean(grayImage(:, column1:column2), 2);

To get a mean horizontal profile between two rows:

horizontalProfile = mean(grayImage(row1:row2, :), 1);

[Lorenzo Gentili]

Thank you for your answer.

[YOGITAA YOGITAA]

could ypu please share your code, i have the same problem

[Image Analyst]

@YOGITAA YOGITAA, here is the code:

https://www.mathworks.com/matlabcentral/answers/389889-how-to-plot-intensity-profile-of-an-image#comment_737092

verticalProfile = mean(grayImage(:, column1:column2), 2);
horizontalProfile = mean(grayImage(row1:row2, :), 1);

Answer 4

[r r]

I have image of a tumor in it and I want to do it the same way https://www.mathworks.com/matlabcentral/answers/389889-how-to-plot-intensity-profile-of-an-image

, but every time the response is

'improfile' requires Image Processing Toolbox.

Error in line (line 4)

c = improfile(I,x,y);

[Walter Roberson]

You need the Image Processing Toolbox to use improfile()

improfile() uses linspace() to create vectors of coordinates between the two endpoints, and uses interp2() to interpolate at the coordinates; it does the interpolation once for each color channel.

The question then becomes what coordinates you should interplate at

Answer 5

[fransec]

Hello everybody,

Anyone knows how to make an image profile like ones you showed, but but along the diagonal of an image?

Thank you so much.

[Walter Roberson]

[rows, cols, ~] = size(YourImage);
n = min(rows,cols);
prof_diag = improfile(YourImage, [1 n], [1 n]);    %diagonal
prof_antidiag = improfile(YourImage, [n 1], [1 n]);  %anti-diagonal

[Christopher Gordon]

Hey! Question along the same lines. I have 30 vertical lines across a 1024 column image/matrix. Is there a way that I can interpolate/extrapolate the remaining columns to get the entire image/matrix?