When Python came to repair the artifact, it was found to be really easy to use!!

author | Junxin

source |  About data analysis and visualization

Today, I want to share with you Python Specific applications in image processing , Since it is image processing , It must be mentioned opencv Module , This module supports many algorithms related to computer vision and machine learning , And the application field is expanding , There are roughly the following areas

• Object recognition ： Judge objects through vision and internal storage

• Image segmentation

• Face recognition

• Safe driving

• human-computer interaction

• wait

Of course, this little editor doesn't intend to put such profound content , Today, let's start with the most basic opencv The module starts with the basic operation of images

Module installation

We install modules through pip By order

pip install opencv-python
pip install opencv-contrib-python

Various operations of pictures

Having learned the positive matrix of linear algebra is not strange . Images are essentially matrices , Grayscale image is a common matrix , And color image is a multidimensional matrix , Our operations on images can be naturally transformed into operations on matrices

Read images

First, let's read the image , It's called cv2.imread() Method , Its syntax is as follows ：

cv2.imread(filename, flag=1)

Among them flag Parameter is used to set the format of the read image , The default is 1, In accordance with RGB Three channel format to read , If I set it to 0, It means to read in the way of gray-scale image single channel ,

import cv2
import numpy as np
img=cv2.imread('1.jpg', 0)

Show the image

After reading the picture , We hope to show it , The function method used here is cv2.imshow(), Its syntax format is as follows

cv2.imshow(name, img)

The parameters are explained as follows ：

• name: It represents the name of the display window

• img: Matrix form of picture

We try to show the pictures read above , The code is as follows

cv2.imshow("grey_img", img)
##  If used cv2.imshow() function , The following must be followed by a function that destroys the window
cv2.destroyAllWindows()

When we run the above code , It can be found that the picture pops up in an instant , But we closed it directly before we could see the picture clearly , The reason lies in cv2.imshow() Function method has no function of delay , We add a delay function , The code is as follows ：

import cv2
import numpy as np
img = cv2.imread('1.jpg')
cv2.imshow("grey_img", img)
cv2.waitKey(0)
cv2.destroyAllWindows()

output

pictures saving

Finally, we save the picture , The function used here is cv2.imwrite(), Its syntax is as follows

cv2.imwrite(imgname, img)

The parameters are explained as follows ：

• imgname: The name of the picture to save

• img: Matrix form of picture

The sample code is as follows

import cv2
import numpy as np
img = cv2.imread('1.jpg')
cv2.imshow("grey_img", img)
cv2.waitKey(0)
cv2.imwrite('1.png', img)
cv2.destroyAllWindows()

Various properties of pictures

Sometimes we want to know the pixel size of the picture , The essence of pictures is matrix , For example, a 1024 Pixels *960 Pixel image , It means that the number of rows in the matrix is 960 That's ok , The number of columns is 1024 Column , stay opencv Called in the module shape() Function method , The code is as follows ：

import cv2
img = cv2.imread('1.jpg')
print(img.shape[0]) #  Row number
print(img.shape[1]) #  Number of columns
print(img.shape[2]) #  The channel number 

output

308
340
3

You can see that the pixels of this picture are 340*380, The number of channels is 3, For gray-scale images , Let's take a look at the properties of the image , The code is as follows ：

img = cv2.imread('1_grey.png', 0)
print(img.shape)

output

(308, 340)

You can see that for gray-scale images , We don't see the number of channels , Only the number of rows and columns

Basic operation of image

Finally, let's do some basic operations on the image , It's nothing more than changing some pixel values , We import a blank picture , By modifying the pixel value in the middle, add a black dot , The code is as follows

import cv2
import numpy as np
img = cv2.imread('2.jpg')
(x, y, z) = img.shape
for i in range(-10, 10):
    for j in range(-10, 10):
        #  Change the pixel value at the right center of the picture ,
        img[int(x/2) + i, int(y/2) + j] = (0, 0, 0)
cv2.imshow("img", img)
cv2.waitKey(0)
cv2.destroyAllWindows()

output

Looking back

It's too voluminous ！AI High accuracy of math exam 81%

This Python Artifact can let you touch fish for a long time ！

2D Transformation 3D, Look at NVIDIA's AI“ new ” magic ！

How to use Python Realize the security system of the scenic spot ？

 Share
Point collection
A little bit of praise
Click to see