So, I have a numpy ndarray with dimensions (984, 1977, 2). What I want to accomplish is to have a numpy ndarray where I do basic math on the final values. So let's say data is my ndarray. And data[0][0] equals to [72 46]. So I want to perform (72 - 46) / (72 + 46) and store that value in my new ndarray for each of the pair. Basically, it represents two stacked bands and the final output needs to be a resultant of the two based on the above formula (not NDVI). I have been looking at numpy's tutorial to find an answer but no luck so far.
For that array
x = data[:,:,0]
y = data[:,:,1]
are the two 'last' column values
res = (x-y)/(x+y)
should be the result you want for all pairs, a (984, 1977) shape array.