Hello Abouzahir,

Pixels in multispectral/hyperspectral images are vectors with n elements (n= number of bands). In my opinion you are right. Each pixel is defined by its location (line and column) and by the value for each one of the bands.

Are you trying to obtain the value for the different bands for each pixel?  

Thanks sir for your answer,

actually i'm working on remote sensing so getting the value of each band is a necessity, so could please tell me how is that done ? and is there any book can help me to understand how multi/hyper images are manipulated.

If I were you, I would create a python script to extract the pixel values of your selected locations. Regarding the book, I recommend "Imaging spectrometry: Basic Principles and prospective applications" Edited by van der Meer and de Jong. It helped me very much.

If you need more info, please do not hesitate in ask me ;) 

Thanks sir,

python script seems to be more comfortable to me.

i appreciate your help Mr.Juan

You`re welcome! Feel free to ask if you need more help. I am used to using arcPy so I can help you if you need it ;)

Also is necessary to know about interleaving in your remote sensing data. There are 3 types of that:

- Band-Sequential Band Interleaved (BSQ)

- Band-Interleaved-by-Pixel Pixel-Interleaved (BIP)

- Band-Interleaved-by-Line Row-Interleaved (BIL)

So, if M is columns number, N- is rows number, Z - is bands number then

BSQ matrix will be M x N x Z

BIP matrix will be Z x M x N

BIL matrix will be M x Z x N