[Opencv Actual Combat] Control the image threshold to achieve binarization and replace the color of the image

Opencv+python realizes binarization of pictures (black and white pictures), and then realizes changing the specified color in the picture, which is classic and effective!

learning target:

When we use opencv to realize image processing, picture binarization and color replacement are two processing methods with relatively high frequency. As a calligraphy lover, I will take processing a pair of calligraphy pictures as an example to realize binarization and color replacement. Color Replacement:

---

Learning Content:

First prepare a picture that needs to be binarized:

we want to replace all the rice-character grids and the beige background with pure white for this exercise, so that we can get a beautiful and simple work. The binarization processing code is as follows:

import cv2 as cv



def load_image():
    src=cv.imread('test.jpg')
    h,w=src.shape[:2]
    #src=cv.resize(src,(w*2,h*3))
    return src
#全局阈值
def threshold_demo(image):
    gray=cv.cvtColor(image,cv.COLOR_BGR2GRAY)
    #为0则使用自动寻找阈值的选项 使用自动寻找即用|分开。
    #如下100为阈值，低于100的像素点全部置为255（即白色）
    ret,binary=cv.threshold(gray,100,255,cv.THRESH_BINARY)#|cv.THRESH_TRIANGLE
    print(ret)
    cv.imshow('change',binary)
	
	#保存二值化后图片change.jpg
    cv.imwrite('change.jpg', binary, [int(cv.IMWRITE_JPEG_QUALITY), 70])
    cv.waitKey(0)
    cv.destroyAllWindows()

threshold_demo(load_image())

Processed picture:

---

Color Replacement:

Here we take replacing the color of the stamp as an example, let black be replaced by red:
first prepare a black and white picture:

color replacement code:

#准备好需要替换颜色的图片yinzhang.jpg
img_rgb = cv2.imread('yinzhang.jpg')

Conv_hsv_Gray = cv2.cvtColor(img_rgb, cv2.COLOR_BGR2GRAY)

res, mask = cv2.threshold(Conv_hsv_Gray, 0, 255,cv2.THRESH_BINARY_INV |cv2.THRESH_OTSU)

#置换颜色[55, 67, 188]为需要置换的红色
#255为需要保留的颜色（即不被替换的颜色），利用掩膜思想覆盖黑色部分
img_rgb[mask == 255] = [55, 67, 188]

cv2.imshow("imgOriginal", img_rgb) # show windows
cv2.imwrite('change.jpg', img_rgb, [int(cv2.IMWRITE_JPEG_QUALITY), 70])

cv2.imshow("output", res) # show windows

cv2.imshow("mask", mask) # show windows

cv2.waitKey(0)

Replacement result:

---

In this way, we realize the segmentation of the threshold and the replacement of the target color