<>python numpy Module detailed explanation and application cases
Introduction to blog
This blog introduces 3 It's a common one numpy method （array,linspace,arange）, At the same time, each method will have an application example , Through the application of examples to better understand and use these methods
* of course numpy There are also some mathematical operations , This is the most familiar one :
sin(), cos(), log(),tan(),exp(),arctan(),arcsin(),arccos(),cot() wait
If you know mathematics, you must know those methods , So I will not introduce it here
<> one , install numpy
windows system ：
1.windows + R Open the console ;
2.pip install numpy.
The results are as follows ：
<> two , The most common method
1.1 array（numpy.array） function
Generate array （ It can be a one-dimensional array , Two dimensional array , Arrays of even higher dimensions ）
The difference between array and list ：
Arrays must be of the same data type ;
The list can be of different data types .
In other ways, the two are similar .
give an example ：
import numpy as np """ The most common operation in the industry , Declare an alias np """ array_0 = np.array([0, 1, 2, 3, 4]
) # One dimensional array array_1 = np.array([[0, 1, 2, 3],[4, 5, 6, 7]]) # Two dimensional array array_2 = np.
array([[[0, 1, 2],[3, 4, 5]],[[6, 7, 8],[9, 10, 11]]]) # 3D array # More dimensions are rarely used
""" Notice here ： Arrays must be made up of elements of the same data type , therefore , When the data types of elements are different , There will be forced replacement !! The priority of quasi exchange is ： character string > Floating point number > integer
For example, the following example ： """ array_3 = np.array([1, 2.23, 4, 5]) # Is cast to floating point array_4 = np.array(
[1, 2.23, 4, 5, 'hello world !']) # Is cast to a string # Print results print(array_0) print(
array_1) print(array_2) print(array_3) print(array_4)
The results of the operation are as follows （output）：
[0 1 2 3 4] # array_0 [[0 1 2 3] [4 5 6 7]] # array_1 [[[ 0 1 2] [ 3 4 5]] [[ 6
7 8] [ 9 10 11]]] # array_2 [1. 2.23 4. 5. ] # array_3 ['1' '2.23' '4' '5'
'hello world !'] # array_4
In short, that is to say ,numpy.array() Is to generate an array , Its type is ：<class ‘numpy.ndarray’>
（ Can pass type（） Method to view ）.
in addition , Subtract a number from an array , for example ：
array_0 = array_0 - 100 # One hundred for each element
It means that every element in the array is subtracted 100!
1.2 array Application examples of the method （ Can handle images ）
import numpy as np import matplotlib.pyplot as plt #
This is a special drawing module , The introduction of this module will be described in another blog post """ It is also a commonly used nickname in the industry """ # Convert the picture to an array （ 3D array ） array_img =
plt.imread('./ Picture name ') """ for example ： array_img =
plt.imread('./0065ErDtgy1geehlwnfb6j30p018g435.jpg') """ # It is equivalent to entering the location of an image file #
The result is a three-dimensional array # adopt print see print(array_img) # print(type(array_img)) # Use the drawing module to show it
plt.imshow(array_img) # Implementation of an array minus a number ： plt.imshow(array_img - 100) # It's OK to add a number plt
.imshow(array_img + 200)
The results of the operation are as follows （output）：
I have already used a picture here , The original picture is as follows , For the effect of running, please see below ：
Picture here ：
1, Corresponding to array_img ： ( No change )
2, Corresponding to array_img - 100 ：（ I made my lovely little sister ugly , It's my fault , Wuwuwu ~~~）
3, Corresponding to array_img + 200 ：（ I made my little sister ugly again , Wuwuwu ~~~, I'm sorry first ~~, But it seems better than the previous one ~）
So , In fact, those software is to take pictures after the transformation into an array , Then through a specific algorithm to achieve beauty , Slimming and other functions .
such as ：
For the commonly used thin face function , We can ：
1, photograph ;
2, Face recognition ;
3, Turn the whole picture into an array ;
4, A specific algorithm is used to deal with the array region corresponding to the face ;
5, last , Display the processed image （ Namely , Picture after face thinning ）.
（ In this case, we don't need to implement this kind of complex function , Those who are interested can explore it by themselves ~）
2.1 linspace Function of method
linspace Method is used to generate a one-dimensional array , The concrete implementation is the interval from the start position to the end position , Produce a specified number of numbers evenly , And these numbers form a one-dimensional array .
import numpy as np array_0 = np.linspace(-10, 20, 30) # The first parameter ： Starting position # The second parameter ： It's over
# The third parameter ： The size of the resulting one-dimensional array , That is, the number of digits specified print(array_0)
Output results ：（ Equivalent to arithmetic sequence !!）
[-10. -8.96551724 -7.93103448 -6.89655172 -5.86206897 -4.82758621 -3.79310345 -
2.75862069 -1.72413793 -0.68965517 0.34482759 1.37931034 2.4137931 3.44827586
4.48275862 5.51724138 6.55172414 7.5862069 8.62068966 9.65517241 10.68965517
11.72413793 12.75862069 13.79310345 14.82758621 15.86206897 16.89655172
17.93103448 18.96551724 20. ]
（ share 30 A number ）
2.2 linspace Application examples of the method
Here is an example of drawing an image , Draw a three-dimensional curve
""" Draw a three-dimensional curve The curve drawn is an equidistant helix Isometric spiral drawing code and the results are as follows ： """ import numpy as np import
matplotlib.pyplot as plt # Import library functions fig = plt.figure() # Establish a three-dimensional coordinate system ax1 = plt.axes(
projection='3d') # 3D drawing z = np.linspace(-5, 5, 50) # use linspace Method definition z coordinate """
from -5 reach 5 The interval of , Equally divided into 50 Point drawing is carried out for each copy Point drawing !! """ x = 5 * np.sin(z) y = 5 * np.cos(z) # definition x And y Coordinates of
ax1.plot3D(x, y, z, 'gray') # Realize the drawing of three-dimensional drawing plt.show() # Displays the result of the drawn image
The results are shown below ：
That's all linspace A brief introduction of the method .
<>3, arange method
3.1 arange Function of method
Generate a starting position , End at end position , An array with a specific step size as the difference value （ It is also equivalent to arithmetic sequence ）：
import numpy as np array_0 = np.arange(-10, 10, 2) # use arange method # The first parameter ： Starting position #
The second parameter ： Termination position # The third parameter ： step # Printing print(array_0)
The operation results are as follows ：
[-10 -8 -6 -4 -2 0 2 4 6 8]
3.2 arange Application examples of the method
Here we draw an image of a surface in three dimensions , The specific code implementation is as follows ：
""" The first four lines are the same as the previous example !! """ import numpy as np import matplotlib.pyplot as plt fig
= plt.figure() ax3 = plt.axes(projection='3d') # definition x coordinate ---> Generate array x = np.arange(-5,
5, 0.1) # definition y coordinate ---> Generate array y = np.arange(-5, 5, 0.1) # ad locum , Generate a two-dimensional array to store coordinates （ Storage of two dimensional coordinates
!!） # （meshgrid It's another way , I won't introduce it here , Interested can refer to the relevant information to learn .） X, Y = np.meshgrid(x, y) # Z
The function relation of Z = 10 * np.log(1000 - X ** 2 - Y ** 2) #
Display image ,cmap Parameters are used to style a drawing ,rainbow It's like a rainbow ax3.plot_surface(X, Y, Z, cmap='rainbow') #
Don't forget plt.show()!! plt.show()
The running results are shown as follows ：
in summary ：
numpy（np） modular That's it ~~~
Of course, the above introduction is not comprehensive enough , Want to really master these modules , Or should we practice more by ourselves , Multi practice , In order to enhance their skills , This article is for your reference only ~
If you like, you can like it （ kiss you ）, Of course, if you don't like it, I hope you don't step on it ~