OpenCV: read and write pictures

Article Directory

1. Read the picture

cv2.imread(filename, flags):

-filename: 文件名称
-flags: 0 读入灰度图片，1 读入彩色图片

cv2.imshow(winname, mat):

 -winname: 窗口名字
 -mat: 要展示的图片矩阵

cv2.waitKey(0): Pause the program so that the picture can be displayed

import cv2
img = cv2.imread("img.jpg", 1)
cv2.imshow('img', img)
cv2.waitKey(0)

Insert picture description here

2. Picture writing

There are four steps to read pictures:

1.文件的读取
2.封装格式解析(jpg,png等格式)
3.数据解码
4.数据加载

The picture read in this way is the original data of the picture

cv2.imwrite(filename, img):

-filename: 图片名称
-img: 图片数据

import cv2
img = cv2.imread("img.jpg", 1)
cv2.imwrite('img1.jpg', img)

True

2.1 Image quality

1. jpgPicture files are compressed at the expense of picture quality, which belongs to lossy compression

cv2.IMWRITE_JPEG_QUALITY: Indicates the current image quality, the compression range is 0-100, and different compression ratios correspond to different image sizes. Let's experience it below:

import cv2
img = cv2.imread("img.jpg", 1)
cv2.imwrite('imgTest.jpg', img, [cv2.IMWRITE_JPEG_QUALITY, 0])

True

The original image is as follows:
Insert picture description here

Image size is 400kb

The compressed picture is:

Insert picture description here

The image size is 40kb, and there is a serious mosaic phenomenon

import cv2
img = cv2.imread("img.jpg", 1)
cv2.imwrite('imgTest2.jpg', img, [cv2.IMWRITE_JPEG_QUALITY, 50])

True

The compressed picture is:

Insert picture description here

The picture size is 200kb, the mosaic phenomenon is not so serious

2. pngThe compression of the image format is lossless, and the transparency can be set

import cv2
img = cv2.imread("img.jpg", 1)
cv2.imwrite('imgTest1.png', img)

True

import cv2
img = cv2.imread("img.jpg", 1)
cv2.imwrite('imgTest2.png', img, [cv2.IMWRITE_PNG_COMPRESSION, 0])

True

The compressed image is:

Insert picture description here

The picture size is: 5.92MB

import cv2
img = cv2.imread("img.jpg", 1)
cv2.imwrite('imgTest3.png', img, [cv2.IMWRITE_PNG_COMPRESSION, 50])

True

The image size becomes 2.34MB

It can be seen that jpgthe lower the value of the pngimage file image quality, the higher the compression ratio, the lower the value of the image file image quality, the lower the compression ratio

3. Pixel

像素: Refers to the small squares that make up the image

RGB: Each color is a combination of RGB (red, green, blue) three colors

颜色深度: For example, 8bit means that the range of each color is 0~255, so there are 256^3 colors in total

图像宽高: Indicates the number of pixels in the horizontal and vertical directions

Calculation method for uncompressed pictures: w * h * 颜色通道(3) * 8 bit / 8(B)

3.1 Pixel reading and writing

Each pixel is composed of 3 parts. Generally, the format of image storage is RGB, but the image read by opecv is in BGR format

We turn the column in the upper left corner of the picture blue

import cv2
img = cv2.imread('img.jpg', 1)
(b, g, r) = img[100, 100] # 读取像素值
print(b, g, r)
#10,100 --- 110, 100
for i in range(1, 1000):
    img[10 + i][100] = (255, 0, 0)
cv2.imshow('imageBlue.png', img)
cv2.waitKey(0)

Insert picture description here