cv2 - Masking image with noise

Question

I have a cv2 image with shape (448,448,3) and want to mask all pixels rgb > 250 with noise. I am doing pixel by pixel now, but it is not efficient. original masked . How can I do such operation in one pass?

rand_mask = np.round(np.random.rand(448, 448, 3) * 255).astype(np.uint8)
    for i in range(0, 448):
        for j in range(0, 448):
                if arr[i, j, 0] and arr[i, j, 1] and arr[i, j, 2] > 250:
                    arr[i, j, :] = rand_mask[i, j, :]

Answer 1

idx = arr>250
arr[idx] = rand_mask[idx]