Mybe bug in OpenCV algorithm for method cv2.erode()

Python
1

Hello Guys i found something anomaly in openCV for method cv2.erode()

i have matrix 4x4 like this

[1,0,0,0],
[1,1,1,0],
[1,1,1,0],
[0,1,1,0],

with Kernel =
[0,1,0],
[1,1,1],
[0,1,0],

when i am try to calculate using manual method in erotion theory i found result equal
[[0 0 0 0]
[0 0 0 0]
[0 1 0 0]
[0 0 0 0]]

but when i am using OpenCV cv2.erode with padding 0 using borderType = cv2.BORDER_CONSTANT and iteration = 1
the result of openCV equal =
[[0 0 0 0]
[1 0 0 0]
[0 1 0 0]
[0 0 0 0]]
its make my value pixel = 1 in (1,0) position but when i am using manual calculation base on algorithm concept its should be = 0 in (1,0) position

can someOne Explain to me why?

2

MRE required.

when asking for BORDER_CONSTANT, you also have to pass a borderValue. the default does not appear to be sensible in this case.