cv2.imread接收cv2.IMREAD_GRAYSCALE作为参数时数组的不同长度

Question

Training_Data=[]
IMG_SIZE=100

def build():
    Directory="D:\projects\Machine_learning\Dog_Cat\PetImages"
    CATEGORY=["Cat","Dog"]
    for category in CATEGORY:
        path=os.path.join(Directory,category)
        class_num=CATEGORIES.index(category)
        for img in tqdm(os.listdir(path)):
            try:
                img_array=cv2.imread(
                    os.path.join(path,img),
                    cv2.IMREAD_GRAYSCALE
                )
                new_array=cv2.resize(img_array(IMG_SIZE,IMG_SIZE))
                Training_Data.append([new_array,class_num])
            except Exception as e:
               pass

When I don't pass cv2.IMREAD_GRAYSCALE, it gives different length of array:

img_array=cv2.imread(os.path.join(path,img))

X = 74598

Y = 24886

Why it is appending 3 times more element in training_data when cv2.IMREAD_GRAYSCALE is not used?

Answer 1

By default, when you don't pass cv2.IMREAD_GRAYSCALE as argument, cv2.imread will read your image with three channels (it will load your image as RGB). When you pass cv2.IMREAD_GRAYSCALE as argument, your amount of bits per pixels is changing (you'll only need 1*8 bit per pixel instead of 3*8).

If you want, you can check out your image depth and channels using depth() and channels() functions, it might help you to better understand what is happening.