Displaying Multispectral Image in MATLAB

I have multispectral image (data3.tif) which is raw image of 6 bands.

I used following code to read the image and matlab reads it properly no issues. I tried to seperate band 4 from raw image and displayed single band it is also works fine.

A=imread('data3.tif');

% b1=im2single(A(:,:,4));

but imshow(A) and imtool(A) doesnot works

i tried multibandread also it does not works.

Actually I need to train neural network for segmentation which accepts [256 256 6] as an input size. To do that first i need to display image (data3.tif) for cliping small patches from the raw data. My problem is imshow and imtool is shows an error if i use imshow(A) and imtool(A).

Question is

how To display multispectral image which has 6 bands in a single image ?

how to crop small patches from raw image?

If i cropped small patch in raw image whether it has same 6 bands?

link for image size 22MB

https://drive.google.com/file/d/1p6ZPOe1nVeQ_COiJHiDXmxFi0X3ezhf-/view?usp=sharing

Answer 1

[Subhadeep Koley]

You can not use imshow(_) or imagesc(_) to display Multi-plane images having more than 3 channels.

You can use sliceViewer(_) to browse through individual plane of your image.

cube = imread('data3.tif');
figure
sliceViewer(cube);
colormap parula

Moreover, if you have the wavelength corresponds to each plane of your multispectral image, you can construct a hypercube(_) object and use the colorize(_) method to generate RGB, CIR, or False-colored image. Also, you can crop your region of interest. Like below,

% Construct the hypercube object
hCube = hypercube('data3.tif', wavelength); % Here wavelength is a C-element vector. C is the number of bands

% Compute RGB image
rgbImg = colorize(hCube, 'method', 'rgb', 'ContrastStretching', true);

% Crop region of interest
row = 100:250;
column = 200:400;
bands = 1:6;
newHcube = cropData(hCube, row, column, bands);

NOTE: The above mentioned fetures come under Image Processing Toolbox's Hyperspectral Imaging Library support package, and can be downloaded from here. For more information on Hyperspectral Imaging Library see the documentation.