How to use numpy array with Image google vision

Question

I want to develop an text_detect application:

image_path = "./image.png"
original_image = cv2.imread(image_path)

crop = take_crop(original_image)