Fix order kwarg for some versions of numpy.copy

It seems it isn't always a valid kwarg. This bypasses that problem
by calling `np.array` directly (which is just what `np.copy` does
under the hood anyway).

skimage/util/arraypad.py

@@ -1532,7 +1532,7 @@ def crop(ar, crop_width, copy=False, order='K'):
     crops = _validate_lengths(ar, crop_width)
     slices = [slice(a, ar.shape[i] - b) for i, (a, b) in enumerate(crops)]
     if copy:
-        cropped = np.copy(ar[slices], order=order)
+        cropped = np.array(ar[slices], order=order, copy=True)
     else:
         cropped = ar[slices]
     return cropped