Arithmetic within numpy arrays

Question

I have read 3 raster images of equal shape (500 by 500) as numpy array, and have put them in this way:

rasters = np.array(A,B,C)

Where A, B, C are 2d numpy arrays belonging to each image.

Now I have to calculate the followings:

result1 = B-A
result2 = C-B

Then,

final_result = np.max([result1,result2],axis = 0)

The final_result should have the same shape of A or B or C (i.e., 500 by 500)

How can I do it?

Is the issue getting A.B and C out of rasters, or doing the subtractions, or something else? — Scott Hunter
– Scott Hunter, Commented Dec 21, 2014 at 9:45

Jaime · Accepted Answer · 2014-12-21 10:08:08Z

3

You can use np.diff and np.max:

np.max(np.diff(rasters, axis=0), axis=0)

Alternatively:

np.max(rasters[1:] - rasters[:-1], axis=0)

answered Dec 21, 2014 at 10:08

Jaime

67.7k19 gold badges128 silver badges164 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

Scott Hunter · Accepted Answer · 2014-12-21 09:48:29Z

1

B-A is accomplished using np.subtract(B,A).

answered Dec 21, 2014 at 9:48

Scott Hunter

50k12 gold badges64 silver badges107 bronze badges

In my real problem, A, B, C...are many. So looking for pythonic way of doing it.

And, I want to do it from the 'rasters'.

@simen: So "I have read 3 raster images" was not exactly true, nor was "The issue is doing subtractions". Good to know.