Converting a jpg image (RGB) into a 2-d array of numbers (2D)

I want to read in an image (a .jpg image) into MATLAB, but using the

Image = imread('image.jpg');

function converts the image into a 3-d or maybe 5-d array i.e. x,y,RGB or x,y,R,G,B. What I want is to be able to read the image into MATLAB as a 2-d array of numbers, so instead of each pixel having 3 numbers to define it's colour (i.e. R G B) it would have just one number to define colour. Then I want to use the

imagesc(Image);

function to view the 2-d array graphically.

Any help with this would be much appreciated.

Thanks,

Terry.

Answer 1

[KIRAN kumar]

I=imread('inputImage.extenstion') 
imshow(I) 
title('Input Image')
 I1=rgb2gray(I);
 figure,imshow(I1) 
 title('gray converted Image')

[Eric Pleace]

Thanks, just what I was looking for :-)

Answer 2

[Sean de Wolski]

Assuming your image is uint8, in order to have a unique value for each pixel color combination, you'll need 256^3 (~16.8million) values. These values will have to be stored in floating point or uint32 values and your images will be much larger. You'll then have to make a colormap of [~16.8million x 3] floating point values to represent the colors you want. This approach, to word it correctly, is stupid. What is so wrong with reading in a 3-dimensional matrix? You could always store it as a vector and reshape it after.

my $0.02.

(unless you just want a gray scale image in which case doc rgb2gray)

[Terence Britton]

The image is an image of the earth, colour scaled for temperature, much like the images on google if you search 'earth temperature' (http://climate.uvic.ca/climate-lab/front_page_pics/temperature.lrg.jpeg).

What I am trying to do is use this image in a simulated interferometry experiment. My demonstrator at Uni has told me that the software we will use to do this, takes in as the input an array of numbers. So I am trying to get the image into a 2-d array of numbers. They don't need to represent the actual temperature numbers, but they do need to be accurate relative to each other. I'm not sure how I'd do this since MATLAB would need to know the colour scale of the image. I do have the colour scale but don't know how to tell MATLAB to look at the scale and turn each colour into the number it represents.

[Sean de Wolski]

rgb2gray()

[KIRAN kumar]

I=imread('inputImage.extenstion')
imshow(I)
title('Input Image')
I1=rgb2gray(I);
figure,imshow(I1)
title('gray converted Image')

[M Shajeeh Mustafa]

I needed the 2-D array to binarize my image using imbinarize(); function. It gives error if your image is a 3-D array

[Walter Roberson]

imbinarize() is not suitable for nonlinear measurements such as temperature represented as color.

If you have data that is linear in some property, you need to convert the 3D data to be in terms of that property, after which you can imbinarize()

For example if you are using color images that are linear in brightness, then you can use rgb2gray() and imbinarize() the result.

Answer 3

[Walter Roberson]

rgb2ind() to get out an array of indices and the colors associated with each index. Then translate each of the colors once through the color scale, creating a vector of temperatures. Take the array of indices you got from rgb2ind, add 1 to it (because indices start from 0) and use that array to index the vector of temperatures to get out an array of temperatures.

Answer 4

[dhia jamaa]

I = imread('image.jpg');
image=mat2gray(I);

[Walter Roberson]

The question turned out to be able converting the colors to temperature values. As the mapping of brightness of color to temperature was not necessarily linear, the color itself needed to be looked up in a conversion table to get temperature. The easiest way to do that is to use rgb2ind() to get a color index for each point, that could be used to index into a vector of conversions of color to associated temperature.