Detailed explanation of the basic operation methods of Python-OpenCV_python

The editor below will share with you a detailed explanation of the basic operation methods of Python-OpenCV. It has a good reference value and I hope it will be helpful to everyone. Let’s follow the editor to take a look

Basic properties

cv2.imread (file name, properties) Read in the image

Attribute: Specify how the image is read from the file

cv2.IMREAD_COLOR: Read in color images, default parameters, Opencv reads color images in BGR mode! ! ! Note

cv2.IMREAD_GRAYSCALE: Read in grayscale images.

cv2.imshow(window name, image file) Display image

Can create multiple windows

cv2.waitKey() keyboard binding function

The function waits for a specific number of milliseconds to see if there is input from the keyboard.

cv2.namedWindow(window name, attribute) Create a window

Attribute: Specify window size mode

cv2.WINDOW_AUTOSIZE: Automatically based on image size Create size

cv2.WINDOW_NORMAL: The window size can be adjusted

cv2.destoryAllWindows(window name) Delete any created window

Code example:

import cv2
 img=cv2.imread('test.py',cv2.IMREAD_COLOR)
 cv2.namedWindow('image',cv2.WINDOW_NORMAL)
 cv2.imshow('image',img)
 cv2.waitKey(0)
 cv2.destoryAllWindows()

cv2.imwrite(save image name, need to save image) Save image

Code example:

 import cv2
 img=cv2.imread('test.png',0)
 cv2.imshow('image',img)
 k=cv2.waitKey(0)
 if k==27: #等待 ESC 键
  cv2.destoryAllWindows()
 elif k==ord('s') #等待 's' 键来保存和退出
  cv2.imwrite('messigray.png',img)
  cv2.destoryAllWindows()

Some operations on images

0x01. Get image attributes

import cv2
img=img.imread('test.png')
print img.shape
#(768,1024,3)
print img.size
#2359296 768*1024*3
print img.dtype
#uint8

0x02. Output text

When processing pictures, output some information directly in the form of text. On the picture

cv2.putText(picture name, text, coordinates, text color)

##0x03. Zoom picture

Implement scaling and saving of images, a common operation when using OpenCV. cv2.resize() supports a variety of interpolation algorithms. By default, cv2.INTER_LINEAR is used. The most suitable one for reducing is cv2.INTER_AREA. The most suitable one for enlarging is cv2.INTER_CUBIC or cv2.INTER_LINEAR.

res=cv2.resize(image,(2*width,2*height),interpolation=cv2.INTER_CUBIC)

Or:

res=cv2.resize(image,None,fx=2,fy=2,interpolation=cv2.INTER_CUBIC)

None here should be the size of the output image , because the scaling factor

0x04 is set later. The image translation is

cv2.warpAffine(src, M, dsize[, dst[, flags[, borderMode[, borderValue]]]])

The translation is Change the position of the image. If you want to move in the (x, y) direction and the movement distance is (tx, ty), you need to construct an offset matrix M.

For example, pan the picture (100,50)

 import cv2
 img=cv2.imread('test.png',1)
 rows,cols,channel=img.shape
 M=np.float32([[1,0,100],[0,1,50]])
 dst=cv2.warpAffine(img,M,(cols,rows))
 cv2.imshow('img',dst)
 cv2.waitKey(0)
 cv2.destoryALLWindows()

Where (cols, rows) represents the size of the output image, M is the transformation matrix, 100 represents the offset of x, and 50 represents the offset of y, in pixels.

0x05. Image rotation

In OpenCV, you first need to construct a rotation matrix, which is obtained through cv2.getRotationMatrix2D.

import cv2
img=cv2.imread('test.png',0)
rows,cols=img.shape
#第一个参数为旋转中心，第二个为旋转角度，第三个为旋转后的缩放因子
M=cv2.getRotationMatrix2D((cols/2,rows/2),45,0.6)
#第三个参数为图像的尺寸中心
dst=cv2.warpAffine(img,M,(2*cols,2*rows))
cv2.imshow('img',dst)
cv2.waitKey(0)
cv2.destoryALLWindows()

0x06. Affine transformation

In affine transformation, the original image All parallel lines in are equally parallel in the resulting image. To create the offset matrix, you need to find three points in the original image and their positions in the output image. Then OpenCV provides cv2.getAffineTransform to create a 2*3 matrix, and finally passes the matrix to the function cv2.warpAffine.

import cv2
import matplotlib.pyplot as plt
import numpy as np
img=cv2.imread('test.png')
rows,cols,ch=img.shape
pts1=np.float32([[50,50],[200,50],[50,200]])
pts2=np.float32([[10,100],[200,50],[100,250]])
M=cv2.getAffineTransform(pts1,pts2)
dst=cv2.warpAffine(img,M,(cols,rows))
plt.subplot(121),plt.imshow(img),plt.title('Input')
plt.subplot(122),plt.imshow(dst),plt.title('Output')
plt.show()

##0x07. Perspective transformation Perspective transformation requires a 3*3 transformation matrix. Make sure the straight line is still straight before and after the transformation. Constructing this matrix requires finding 4 points in the input image and their corresponding positions in the output image. Any three of these four points cannot be collinear. Transformation matrix OpenCV provides cv2.getPerspectiveTransform() construction. Then pass the matrix into the function cv2.warpPerspective.

import cv2
import numpy as np
import matplotlib.pyplot as plt
img=cv2.imread('test.png')
rows,cols,ch=img.shape
pts1=np.float32([[56,65],[368,52],[28,387],[389,390]])
pts2=np.float32([[0,0],[300,0],[0,300],[300,300]])
M=cv2.getPerspectiveTransform(pts1,pts2)
dst=cv2.warpPerspective(img,M,(300,300))
plt.subplot(121),plt.imshow(img),plt.title('Input')
plt.subplot(122),plt.imshow(dst),plt.title('Output')
plt.show()

##0x09. Image regions of Interest

Sometimes it is necessary to operate on a specific area of ​​an image, and the ROI is obtained using Numpy index.

import cv2
import numpy as np
import matplotlib.pyplot as plt

image=cv2.imread('test.png')
rows,cols,ch=image.shape
tall=image[0:100,300:700]
image[0:100,600:1000]=tallall
cv2.imshow("image",image)
cv2.waitKey(0)
cv2.destoryALLWindows()

##0x10. Channel split/merge processing

Sometimes it is necessary to operate the three BGR channels separately. At this time, the BGR needs to be split into a single channel. At the same time, sometimes it is necessary to merge independent channel images into a BGR image.

Use OpenCV library function version

import cv2
import numpy as np
import matplotlib.pyplot as plt

image=cv2.imread('pitt1.jpg')
rows,cols,ch=image.shape
#拆分通道，cv2.split()是一个比较耗时的操作。只有需要时使用，尽量Numpy
b,g,r=cv2.split(image)
print b.shape
#(768,1024)
#合并通道
image=cv2.merge(b,g,r)

Use Numpy index version:

import cv2
import numpy as np
import matplotlib.pyplot as plt

image=cv2.imread('pitt1.jpg')
rows,cols,ch=image.shape
#直接获取
b=img[:,:,0]

The above is the detailed content of Detailed explanation of the basic operation methods of Python-OpenCV_python. For more information, please follow other related articles on the PHP Chinese website!