【OpenCV】模板匹配、图像二值化

一、模板匹配

1.1 概念

  **模板匹配：**就是在整个图像区域发现与给与子图像匹配的小块区域。所以模板匹配首先需要一个模板图像T（给与的子图像），另外还需要一个待检测图像-原图像s。
  工作方法： 在待检测图像上，从左到右，从上向下计算模板图像与重叠子图像的匹配度，匹配程度越大，两者相同的可能性越大。

1.2 实战cv.matchTemplate

  cv.matchTemplate(图片，模板，匹配方法)

import cv2 as cv
import numpy as np
from matplotlib import pyplot as plt

def template_demo():
   
    target=cv.imread(r'D:\Project\Opencv\Learning01\angelababy.jpg')
    tpl=cv.imread(r'D:\Project\Opencv\Learning01\muban.png')
    cv.imshow('template',tpl)
    cv.imshow('target',target)
    '''三种匹配时的度量方法'''
    method=[cv.TM_SQDIFF_NORMED,cv.TM_CCORR_NORMED,cv.TM_CCOEFF_NORMED]
    th,tw=tpl.shape[0:2]
    for md in method:
        print(md)
        '''模板匹配'''
        result=cv.matchTemplate(target,tpl,md)#原图像、模板、匹配方法
        min_val,max_val,min_loc,max_loc=cv.minMaxLoc(result)
        if md==cv.TM_SQDIFF_NORMED:
            tl=min_loc
        else:
            tl=max_loc
        br=(tl[0]+tw,tl[1]+th)
        cv.rectangle(target,tl,br,(0,0,255),2)
        cv.imshow('result'+str(md),target)

template_demo()
cv.waitKey(10000)

二、图像二值化

2.1 概念

  二值图像：
  图像中只有两种颜色[0,1]。灰度图像是[0,255]彩色图像是[[0,255],[0，255][0,255]。
                
         

2.2 实战

  二值化之前要先把图片变为灰度图

2.2.1 全局阈值cv.threshold

  ret,binary=cv.threshold()

import cv2 as cv
import numpy as np
from matplotlib import pyplot as plt
''''''
def threshold_demo():
    ''''''
    target=cv.imread(r'D:\Project\Opencv\Learning01\angelababy.jpg')
    """变为灰度图像"""
    gray=cv.cvtColor(target,cv.COLOR_BGR2GRAY)
    """变为二值图像"""
    #########方法1：此时cv.THRESH_OTSU计算阈值，0,255作废
    ret,binary=cv.threshold(gray,0,255,cv.THRESH_BINARY|cv.THRESH_OTSU)#参数：原图像，0,255，二值化方法
    cv.imshow('origin', target)
    print('yuzhi:',ret)#自动计算的分割阈值
    cv.imshow('binary',binary)
    # #########方法2，自动全局阈值，此时cv.THRESH_TRIANGLE计算阈值，0,255作废
    # ret, binary = cv.threshold(gray, 0, 255, cv.THRESH_BINARY | cv.THRESH_TRIANGLE)
    # cv.imshow('origin', target)
    # print('yuzhi:', ret)  # 自动计算的分割阈值
    # cv.imshow('binary', binary)
    # #########方法3，自动指定阈值，小于127是黑色，大于127是白色
    # ret, binary = cv.threshold(gray, 127, 255, cv.THRESH_BINARY )
    # cv.imshow('origin', target)
    # print('yuzhi:', ret)  # 自动计算的分割阈值
    # cv.imshow('binary', binary)
    # #########方法4.取反.大于127是黑色，小于127是白色
    # ret, binary = cv.threshold(gray, 127, 255, cv.THRESH_BINARY_INV)
    # cv.imshow('origin', target)
    # print('yuzhi:', ret)  # 自动计算的分割阈值
    # cv.imshow('binary', binary)
    # #########方法5.截断阈值，大于127变成127，
    # ret, binary = cv.threshold(gray, 127, 255, cv.THRESH_TRUNC)
    # cv.imshow('origin', target)
    # print('yuzhi:', ret)  # 自动计算的分割阈值
    # cv.imshow('binary', binary)
    # #########方法6.小于127变为0
    # ret, binary = cv.threshold(gray, 127, 255, cv.THRESH_TOZERO)
    # cv.imshow('origin', target)
    # print('yuzhi:', ret)  # 自动计算的分割阈值
    # cv.imshow('binary', binary)

threshold_demo()
cv.waitKey(10000)

2.2.2 局部（自适应）阈值cv.adaptiveThreshold

  binary=cv.adaptiveThreshold()

import cv2 as cv
import numpy as np
from matplotlib import pyplot as plt
''''''
def local_threshold():
    ''''''
    target=cv.imread(r'D:\Project\Opencv\Learning01\angelababy.jpg')
    """变为灰度图像"""
    gray=cv.cvtColor(target,cv.COLOR_BGR2GRAY)
    """变为二值图像"""
    #########自适应阈值
    ###方法1：cv.ADAPTIVE_THRESH_MEAN_C
    dst=cv.adaptiveThreshold(gray,255,cv.ADAPTIVE_THRESH_MEAN_C,cv.THRESH_BINARY,25,10)
    cv.imshow('target', target)
    cv.imshow('binary',dst)
    ###方法2：cv.ADAPTIVE_THRESH_GAUSSIAN_C
    # dst=cv.adaptiveThreshold(gray,255,cv.ADAPTIVE_THRESH_GAUSSIAN_C,cv.THRESH_BINARY,25,10)
    # cv.imshow('target', target)
    # cv.imshow('binary',dst)

local_threshold()
cv.waitKey(10000)

2.2.3 自定义阈值cv.threshold

  ret,binary=cv.threshold()
  例使用均值阈值

import cv2 as cv
import numpy as np
from matplotlib import pyplot as plt
''''''
def custom_threshold():
    ''''''
    target=cv.imread(r'D:\Project\Opencv\Learning01\angelababy.jpg')
    """变为灰度图像"""
    gray=cv.cvtColor(target,cv.COLOR_BGR2GRAY)
    """变为二值图像"""
    h,w=gray.shape[0:2]
    m=np.reshape(gray,[1,w*h])
    mean=m.sum()/(w*h)
    print('mean',mean)
    ret,binary=cv.threshold(gray,mean,255,cv.THRESH_BINARY)
    cv.imshow('target',target)
    cv.imshow('binary',binary)


custom_threshold()
cv.waitKey(10000)

三、超大图像二值化

  二值化：全局阈值、局部阈值、自定义阈值。对于超大图像，通常先图像分块，再局部二值化（常用）

import cv2 as cv
import numpy as np
from matplotlib import pyplot as plt
''''''
def big_image_binary():
    ''''''
    big_img=cv.imread(r'D:\Project\Opencv\Learning01\bigimage.png')
    print(big_img.shape)
    #设置分成的小块的大小
    cw=125
    ch=125
    h,w=big_img.shape[0:2]
    gray=cv.cvtColor(big_img,cv.COLOR_BGR2GRAY)#转灰度图
    for row in range(0,h,ch):
        for col in range(0,w,cw):
            roi=gray[row:row+ch,col:col+col+cw]#得到一小块图片
            """计算小块图像的二值化的结果"""
            dst=cv.adaptiveThreshold(roi,255,cv.ADAPTIVE_THRESH_GAUSSIAN_C,cv.THRESH_BINARY,127,20)
            """图像块二值化的结果赋值到原图"""
            gray[row:row+ch,col:col+col+cw]=dst
            print(np.std(dst),np.mean(dst))
    cv.imshow('big binary',gray)


big_image_binary()
cv.waitKey(10000)