OpenCV + matchTemplate 的简单例子
昌正奇
# 导入所需库文件，numpy和cv2。
import cv2
import numpy as np
# 加载原始图像和要搜索的图像模板
img = cv2.imread('wife.jpg')
# OpenCV对原始图像进行处理，创建一个灰度版本
img_gray = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)
# 加载将要搜索的图像模板#记录图像模板的尺寸
template = cv2.imread('template2.jpg',0)
# 记录模板尺寸
w, h, = template.shape[::-1]#输出前两列
# 在灰度图像里进行处理和查找匹配。然后使用相同的坐标在原始图像中进行还原并输出。
# cv2.matchTemplate(image, templ, method[, result])
res = cv2.matchTemplate(img_gray,template,cv2.TM_CCOEFF_NORMED) #TM_CCOEFF_NORMED是标准相关性系数匹配,matchtemplate常用方法
# 设定阈值
threshold = 0.7
# 输出res大于70%的
loc = np.where( res >= threshold)
# 使用灰度图像中的坐标对原始RGB图像进行标记
for pt in zip(*loc[::-1]):
cv2.rectangle(img, pt, (pt[0] + w, pt[1] + h), (7,249,151), 2)
# 显示图像
cv2.imshow('matchTemplate.jpg', img)
# save图像
cv2.imwrite('matchTemplate.jpg', img)
Q&A
Q1:cv2.imread()
A1:
cv2.imread()读取图片,第二个参数：
- cv2.IMREAD_COLOR：加载彩色图片，这个是默认参数。
- cv2.IMREAD_GRAYSCALE：以灰度模式加载图片。
- cv2.IMREAD_UNCHANGED：包括alpha。
上面三个flag分别对应的值为1，0，-1。
Q2:zip函数是什么
A2:
zip使用方法实例
>>>a = [1,2,3]
>>> b = [4,5,6]
>>> c = [4,5,6,7,8]
>>> zipped = zip(a,b) # 打包为元组的列表
[(1, 4), (2, 5), (3, 6)]
>>> zip(a,c) # 元素个数与最短的列表一致
[(1, 4), (2, 5), (3, 6)]
>>> zip(*zipped) # 与 zip 相反，*zipped 可理解为解压，返回二维矩阵式In []: loc[::-1]
Out []:
(array([522, 521, 522, 523, 521, 522, 523, 520, 521, 522, 523, 520, 521,
522, 523, 520, 521, 522, 523, 520, 521, 522, 523, 520, 521, 522,
523, 519, 520, 521, 522, 523, 519, 520, 521, 522, 523, 519, 520,
521, 522, 523, 519, 520, 521, 522, 523, 518, 519, 520, 521, 522,
523, 518, 519, 520, 521, 522, 523, 518, 519, 520, 521, 522, 523,
518, 519, 520, 521, 522, 523, 517, 518, 519, 520, 521, 522, 523,
517, 518, 519, 520, 521, 522, 523, 517, 518, 519, 520, 521, 522,
523, 517, 518, 519, 520, 521, 522, 523, 517, 518, 519, 520, 521,
522, 523, 517, 518, 519, 520, 521, 522, 523, 517, 518, 519, 520,
521, 522, 517, 518, 519, 520, 521, 522, 517, 518, 519, 520, 521,
522, 517, 518, 519, 520, 521, 522, 517, 518, 519, 520, 521, 522,
517, 518, 519, 520, 521, 517, 518, 519, 520, 521, 517, 518, 519,
520, 521, 517, 518, 519, 520, 521, 517, 518, 519, 520, 521, 517,
518, 519, 520, 521, 517, 518, 519, 520, 521, 517, 518, 519, 520,
517, 518, 519, 520, 517, 518, 519, 520, 517, 518, 519, 520, 517,
518, 519, 517, 518, 519, 517, 518, 519, 517, 518, 519, 518],
dtype=int64),
array([834, 835, 835, 835, 836, 836, 836, 837, 837, 837, 837, 838, 838,
838, 838, 839, 839, 839, 839, 840, 840, 840, 840, 841, 841, 841,
841, 842, 842, 842, 842, 842, 843, 843, 843, 843, 843, 844, 844,
844, 844, 844, 845, 845, 845, 845, 845, 846, 846, 846, 846, 846,
846, 847, 847, 847, 847, 847, 847, 848, 848, 848, 848, 848, 848,
849, 849, 849, 849, 849, 849, 850, 850, 850, 850, 850, 850, 850,
851, 851, 851, 851, 851, 851, 851, 852, 852, 852, 852, 852, 852,
852, 853, 853, 853, 853, 853, 853, 853, 854, 854, 854, 854, 854,
854, 854, 855, 855, 855, 855, 855, 855, 855, 856, 856, 856, 856,
856, 856, 857, 857, 857, 857, 857, 857, 858, 858, 858, 858, 858,
858, 859, 859, 859, 859, 859, 859, 860, 860, 860, 860, 860, 860,
861, 861, 861, 861, 861, 862, 862, 862, 862, 862, 863, 863, 863,
863, 863, 864, 864, 864, 864, 864, 865, 865, 865, 865, 865, 866,
866, 866, 866, 866, 867, 867, 867, 867, 867, 868, 868, 868, 868,
869, 869, 869, 869, 870, 870, 870, 870, 871, 871, 871, 871, 872,
872, 872, 873, 873, 873, 874, 874, 874, 875, 875, 875, 876],
dtype=int64))
Q3:cv2.rectangle()
A3:Python: cv2.rectangle(img, pt1, pt2, color[, thickness[, lineType[, shift]]]) → None
感谢以下链接提供方法和解决思路
http://www.runoob.com/python/python-func-zip.html
http://bluewhale.cc/2017-09-22/use-python-opencv-for-image-template-matching-match-template.html
https://docs.opencv.org/trunk/d4/dc6/tutorial_py_template_matching.html
https://docs.opencv.org/2.4/modules/core/doc/drawing_functions.html
http://blog.topspeedsnail.com/?s=template.shape （真的是个很棒的博客，强烈推荐）
类似资料：