The meaning of linetypes enumeration values (line_4, line_8, line_aa) in opencv

stay OpenCV Almost all the functions related to drawing in this LineTypes Parameter settings . Let's say functions line()、 function putText()、 function drawContours()、 function rectangle() etc. .

stay OpenCV4.1.2 In the official documents of ,LineTypes There are the following enumeration values ：

From this we can see that , There are four enumerated values , Respectively FILLED 、LINE_4 、LINE_8 、LINE_AA .
among FILLED There is no explanation for , This indicates that the enumeration value should be useless .

Let's introduce LINE_4 、LINE_8 、LINE_AA The meaning of .

These values are used to determine the line generation algorithm . We know that a line segment is actually composed of many pixels . Take drawing line segments as an example , When the two endpoints of the line segment are determined , In fact, you can determine the coordinates of those points in the middle of the line segment . So how to calculate the coordinates of these points in the middle ？ That is calculated by the relevant algorithm .
stay OpenCV We can use LINE_4 、LINE_8 、LINE_AA To specify which algorithm is used to calculate the coordinates of these points in the middle .
LINE_4 Represents the pixel points on the line segment calculated by the algorithm used , There are only four directions between two adjacent points .
LINE_8 Represents the pixel points on the line segment calculated by the algorithm used , There are only eight directions between two adjacent points .
LINE_AA Represents the pixel points on the line segment calculated by the algorithm used , There are more than eight directions between two adjacent points , For example, sixteen 、 Thirty two or something . How many , You have to look at the source code of the algorithm ,

LINE_AA In the above explanation is “antialiased line”, It means anti huge tooth line .LINE_4 and LINE_8 It belongs to the huge tooth line , Why? LINE_4 and LINE_8 It belongs to the huge tooth line ？ Because their direction is very limited ,LINE_8 There are eight directions ,LINE_4 There are only four directions . Limited directional resolution , The drawn lines naturally look like huge teeth .

Be careful ： To distinguish four connected domains and eight connected domains from the LINE_4 and LINE_8 And difference . The four pixels of a pixel are called its four connected regions , One pixel 、 Next 、 Left 、 Right 、 The northwest 、 The northeast 、 southwest 、 The southeast pixel is called its eight connected region . obviously , If we want to expand the number of angles formed by a pixel and its adjacent points , Then we have to expand Dalian Tongyu , For example, we can call the following point P Of 5×5-1=24 Connected domain , It's just P Point and 24 The points in a connected domain are not necessarily connected .

Next , Look at a sample code , To experience these kinds of LineTypes The difference in effect .

#  Blogger WeChat /QQ 2487872782
#  If you have any questions, you can contact the blogger 
#  If you need image processing development, please also contact the blogger 
#  Image processing technology exchange QQ Group  271891601

import cv2
import numpy as np

str1 = 'WR'
thick = 3
Scale = 6

img_LINE_4 = np.zeros((300, 300, 3), dtype='uint8')
cv2.putText(
    img=img_LINE_4,
    text=str1,
    org=(20, 150),
    fontFace=cv2.FONT_HERSHEY_SIMPLEX,
    fontScale=Scale,
    color=(0, 255, 0),
    thickness=thick,
    lineType=cv2.LINE_4)
cv2.imshow('LINE_4', img_LINE_4)


img_LINE_8 = np.zeros((300, 300, 3), dtype='uint8')
cv2.putText(
    img=img_LINE_8,
    text=str1,
    org=(20, 150),
    fontFace=cv2.FONT_HERSHEY_SIMPLEX,
    fontScale=Scale,
    color=(0, 255, 0),
    thickness=thick,
    lineType=cv2.LINE_8)
cv2.imshow('LINE_8', img_LINE_8)

img_LINE_AA = np.zeros((300, 300, 3), dtype='uint8')
cv2.putText(
    img=img_LINE_AA,
    text=str1,
    org=(20, 150),
    fontFace=cv2.FONT_HERSHEY_SIMPLEX,
    fontScale=Scale,
    color=(0, 255, 0),
    thickness=thick,
    lineType=cv2.LINE_AA)
cv2.imshow('LINE_AA', img_LINE_AA)


cv2.waitKey(0)
cv2.destroyAllWindows()

The operation results are as follows ：

From the running results, we can see that ,LINE_4 and LINE_8 Not much difference , But you can clearly feel LINE_AA Than LINE_4 and LINE_8 Much smoother , That is, good anti aliasing performance .