In python opencv, we can use cv2.matchTemplate() to match images. Here are six template matching algorithms we can use. In this tutorial, we will discuss these algorithms.
OpenCV Template Matching Algorithms
If (x, y) in image1 and (x’, y’) in image2. In order to find the similarity between them, we can use six methods:
| Algorithm | Equation | |
| cv2.TM_CCOEFF |  |
The bigger the better |
| cv2.TM_CCOEFF_NORMED |  |
-1<=R<=1
R=1 is the best |
| cv2.TM_CCORR |  |
R=0 is the worst
The bigger the better |
| cv2.TM_CCORR_NORMED |  |
0<=R<=1
R=0 is the worst |
| cv2.TM_SQDIFF |  |
R=0 is the best |
| cv2.TM_SQDIFF_NORMED |  |
0<=R<=1
R=0 is the best |
How to use these algorithm?
Here is an example:
res = cv2.matchTemplate(img1, img2, cv2.TM_SQDIFF_NORMED)
where res is a numpy.ndarray, the value of it is computed using cv2.TM_SQDIFF_NORMED.