Opencv picture rotation

OpenCV Two functions of image rotation

1. cv2.getRotationMatrix2D Obtain affine change matrix
2. cv2.warpAffine Make affine changes

getRotationMatrix2D Obtain affine change matrix （warpAffine Important parameters of method ）

rot_mat =  cv2.getRotationMatrix2D(center, -5, 1)

Parameter description ：

• center Express The position of the middle point ,
• -5 Express Anti-clockwise rotate 5 degree ,
• 1 It indicates that The zoom
  Return value ：
  rot_mat Affine change matrix
  for example ：
  

warpAffine Make affine changes

img_rotated_by_alpha = cv2.warpAffine(img, rot_mat, (img.shape[1], img.shape[0]))

Parameter description ：

• img Represents the input image ,
• rot_mat Represents an affine change matrix ,
• (image.shape[1], image.shape[0]) Represents the size of the transformed picture ,image.shape[1] Indicates width ,image.shape[0] High

Return value ：
Rotated image

Code instructions ：

First step ： Read in the picture , Show pictures

The second step ： Get the width of the picture 、 Long 、 The channel number , structure [0, 0, w, h] list , Then use this list to materialize a bbox object （ Used to calculate the center point ）

The third step ： Calculating the center value , Substitute it into cv2.getRotationMatrix2D Generate affine change matrix

Step four ： Use cv2.warpAffine(img, rot_mat, (img.shape[1], img.shape[0])) Get the image after affine change

Step five ： If you calculate a point on the source image , In the position of the converted image . If it is a point on the source image , Then the coordinate value after the change is
(rot_mat[0][0] * x + rot_mat[0][1] * y + rot_mat[0][2], rot_mat[1][0] * x + rot_mat[1][1] * y + rot_mat[1][2]

Step six ： Show rotating pictures , In order to ensure , You can also take screenshots here

import cv2
import numpy as np

class BBox(object):

    def __init__(self, bbox):
        self.left = bbox[0]
        self.top = bbox[1]
        self.right = bbox[2]
        self.bottom = bbox[3]


img = cv2.imread('0001.jpg')
cv2.imshow('img', img)   #  Display images 
cv2.waitKey(0)
h, w, c = img.shape  #  The height of the image , wide , passageway ; Be careful ：OpenCV  Get high first ！ Then there is kuanhe channel .
box = [0, 0, w, h]   #  Create a list , common 4 term , The first two terms represent the origin , The third item is width , The fourth item is high 
bbox = BBox(box)     #  establish BBox object 


center = ((bbox.left + bbox.right) / 2, (bbox.top + bbox.bottom) / 2)  #  Calculate the center point 
rot_mat = cv2.getRotationMatrix2D(center, -5, 1)	#  Affine change matrix 
img_rotated_by_alpha = cv2.warpAffine(img, rot_mat, (img.shape[1], img.shape[0]))  
# rot_mat  Affine change matrix ,warpAffine An important parameter of 
# img.shape[1] Is the image width ,img.shape[0] For image height  img by OpenCV Variable 

#  Get the position of key points after image rotation 
# lanmark_ = np.asarray([(rot_mat[0][0] * x + rot_mat[0][1] * y + rot_mat[0][2],
# rot_mat[1][0] * x + rot_mat[1][1] + rot_mat[1][2]) for (x, y) in landmark])

cv2.imshow('img_rotated', img_rotated_by_alpha) #  Displays the rotated image 
cv2.waitKey(0)

# face = img_rotated_by_alpha[bbox.top:bbox.bottom + 1, bbox.left:bbox.right + 1]
#
# cv2.imshow('face', face)
# cv2.waitKey(0)


#  Be careful   The code is in Jupyter Notebook There are bug

original image

Rotate the image

About img.shape[0]、[1]、[2] What does it stand for

img.shape[0]： The vertical size of the image （ Height ）
img.shape[1]： The horizontal size of the image （ Width ）
img.shape[2]： The number of channels in the image

In the matrix ,[0] It means the number of rows ,[1] Indicates the number of columns .