[FIXED] Unexpected behaviour in numpy np.where with np.logical_and

Issue

I want to find an RGB pixel in a numpy tri-dimensional array (X/Y/RGB) created with Pillow

conditions = np.logical_and(np.logical_and(array[:,:,0]==rgb[0],array[:,:,1]==rgb[1]),array[:,:,2]==rgb[2])
res = np.flip(np.transpose(np.where(conditions))).tolist()

It works like a charm.

However, I'd like to add a tolerance, so I just changed the conditions by:

abs(array[:,:,i]-rgb[i]) <= tolerance

But this does not work as expected, it returns zeros instead of pixels in the tolerated range. How can I add tolerance to my upper snippet?

Solution

Problem was calling np.array() on a PIL Image automatically create a dtype=uint8. During my logical tests value can go higher than 255, thus the unexpected behaviour

Logic can be changed to make it work, but array = np.array(im,dtype=np.int16) also does the trick (even if it's less optimized)

Answered By - Turbolay