opencv中模板匹配

函数：cv2.matchTemplate(img, template, method)
原理：模板在原图上滑动，计算模板与图像被模板覆盖的部分的差别程度，若原图形大小是AB，则输出结果矩阵是(A-a+1)(B-b+1)，常见的计算方法有6种：
1）TM_SQDIFF:计算平方不同，计算出来的值越小，越相关
2) TM_CCORR:计算相关性，计算出来的值越大，越相关
3) TM_CCOEFF:计算相关系数，计算出来的值越大，越相关
4) TM_SQDIFF_NORMED:计算归一化平方不同，计算出来的值越接近0，越相关
5) TM_CCORR_NORMED:计算归一化相关性，计算出来的值越接近1，越相关
6) TM_CCOEFF_NORMED:计算归一化相关系数，计算出来的值越接近1，越相关

单个模板匹配代码：

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

# cv2.imread(filename, flag), flag默认为1
# flag = -1, 9位深度，原通道
# flag = 0, 8位深度，1通道（灰度图）
# flag = 1, 8位深度，3通道
# flag = 2, 原深度，1通道
# flag = 3, 原深度，3通道
# flag = 4, 8位深度，3通道
img = cv2.imread('图片路径', 0)
template = cv2.imread('模板路径', 0)
h, w = template.shape[:2]

methods = ['cv2.TM_SQDIFF', 'cv2.TM_CCORR', 'cv2.TM_CCOEFF',
          'cv2.TM_SQDIFF_NORMED', 'cv2.TM_CCORR_NORMED', 'cv2.TM_CCOEFF_NORMED']

# res = cv2.matchTemplate(img, template, cv2.TM_CCORR_NORMED)
# print(res.shape)
#
# min_val, max_val, min_loc, max_loc = cv2.minMaxLoc(res)
# print(min_val,max_val,min_loc,max_loc)

for meth in methods:
    img2 = img.copy()

    # 匹配方法的真值
    method = eval(meth)
    print(method)
    res = cv2.matchTemplate(img2, template, method)
    min_val, max_val, min_loc, max_loc = cv2.minMaxLoc(res)

    if method in [cv2.TM_SQDIFF, cv2.TM_SQDIFF_NORMED]:
        top_left = min_loc
    else:
        top_left = max_loc
    bottom_right = (top_left[0] + w, top_left[1] + h)

    cv2.rectangle(img2, top_left, bottom_right, 255, 2)

    plt.subplot(121), plt.imshow(res, cmap='gray')
    plt.xticks([]), plt.yticks([]) # 隐藏坐标轴
    plt.subplot(122), plt.imshow(img2, cmap='gray')
    plt.xticks([]), plt.yticks([])
    plt.suptitle(meth)
    plt.show()

多个模板匹配代码：

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

img = cv2.imread('图片路径', cv2.COLOR_BGR2GRAY)
template = cv2.imread('模板路径', cv2.COLOR_BGR2GRAY)
h, w = template.shape[:2]

res = cv2.matchTemplate(img, template, cv2.TM_CCOEFF_NORMED)
threshold = 0.8

loc = np.where(res >= threshold)
# *表示可选参数
for pt in zip(*loc[::-1]):
    bottom_right = ((pt[0] + w), (pt[1] + h))
    cv2.rectangle(img, pt, bottom_right, (0, 0, 255), 2)

cv2.imshow('img', img)
cv2.waitKey(0)
cv2.destroyAllWindows()