Why np.ones creates a black image?

Question

I want to understand what is the difference between np.zeros and np.ones . In my bellow code both create a black images. But what is the use of np.zeros and np.ones . When should i use them. What is the purpose. I know np.ones fill matrix with 1 but what is the purpose. What i want to create a white image

img = np.ones((512,512,3), np.uint8)
cv2.imshow('IMG', img)
cv2.waitKey(0)
cv2.destroyAllWindows()

img = np.zeros((512,512,3), np.uint8)
cv2.imshow('IMG', img)
cv2.waitKey(0)
cv2.destroyAllWindows()

Answer 1

Initialize your array as float data type. Only then 0 will mean black and 1 will mean white:

img = np.ones((512,512,3), np.float)

Explanation : When imshow receives an array of ints, it will assume, that 0 is the minimum and 255 the maximum value, representing black and white in a greyscale image. However if it receives an array of floats it sets these values to 0.0 and 1.0, because it assumes a scaled input.

Answer 2

np.zeros "Return a new array of given shape and type, filled with zeros."

np.ones "Return a new array of given shape and type, filled with ones."

np.zeros(10)
Out: array([0., 0., 0., 0., 0., 0., 0., 0., 0., 0.])
np.ones(10)
Out:array([1., 1., 1., 1., 1., 1., 1., 1., 1., 1.])

Answer 3

You need to scale by the number of bits you have in a pixel. For example,

img = 255*np.ones((512, 512, 3), uint8)

will produce a white image for 8-bit pixels.