[python]Determine the brightness of the image

Overview

I was doing image recognition and wanted to judge the brightness of the image and classify it.

Specifically, I'm shooting indoors, but when the light is off (= no people in the room), it's useless to judge considering the nature of the application, so I wanted to avoid entering the judgment process.

So, I thought of a method to judge the brightness of the captured image.

Method

The procedure is as follows.

1. Convert image to grayscale and make it binary

2. Add up all the components in the image

3. Normalize by image size

When the image is grayscale, the values are higher in bright areas and lower in dark areas. By using this, the standardized value by adding all the components in the image can be used for judgment as the "brightness" of the image.

The code is below.

def judgeLight(file):

    img=cv2.imread(file, cv2.IMREAD_GRAYSCALE) #1 load as grayscale
    img=img.astype('float')
    img/=255

    sumGray=img.sum()  #2 add up all the component
    meanGray=sumGray/(img.shape[0]*img.shape[1])  #3　Normalize
    
    return meanGray

I will try it with some images.

imgNames=glob.glob('*.png')

for imgName in imgNames:
    
    m=judgeLight(imgName)
    
    print(meanGray)
    
    plt.figure()
    plt.gray()
    plt.imshow(img)
    plt.show()

You can see that the darker image has a smaller value.

After that, set the threshold appropriately and remove the dark image.

Lastly

I thought about it myself because it didn't come out even if I googled variously, but maybe there is a more proper method.