CODEWITHSUNDEEP
pythonpython-2.7numpyopencv
Programer Beginner
Programer Beginner

Reputation: 1419

Python 2.7 OpenCV matchTemplate, Get only the best match

THRESHOLD = 0.9
results = cv2.matchTemplate(image, template, cv2.TM_CCOEFF_NORMED)
locations = numpy.where( results >= THRESHOLD) # <---

With the code above (where the arrow is) I can get all the matches that are above a certain threshold.

But I want just to get the best match from the results and then check the best match if it is above the threshold or not. How do I do it?

Note: Please make your answer beginner-friendly.

Upvotes: 1

Views: 2307

Answers (1)

Jeru Luke
Jeru Luke

Reputation: 21203

You ought to use cv2.minMaxLoc(). This will return only one and the best match.

min_val, max_val, min_loc, max_loc = cv2.minMaxLoc(results)

top_left = max_loc
bottom_right = (top_left[0] + w, top_left[1] + h)

cv2.rectangle(image.copy(), top_left, bottom_right, 255, 2)

**Illustration: **

I have Messi on the football ground as the images and his face as the template.

img = cv2.imread('C:/Users/selwyn77/Desktop/messi.jpg', 0)
img2 = cv2.imread('C:/Users/selwyn77/Desktop/messi.jpg', 1)
template = cv2.imread('C:/Users/selwyn77/Desktop/messi_face.jpg', 0)
w, h = template.shape[::-1]

#img3 = img2.copy()

res = cv2.matchTemplate(img,template,cv2.TM_CCOEFF_NORMED)
min_val, max_val, min_loc, max_loc = cv2.minMaxLoc(res)

top_left = max_loc
bottom_right = (top_left[0] + w, top_left[1] + h)

cv2.rectangle(img2, top_left, bottom_right, (0, 255, 0), 2)

cv2.imshow('img', img2)

enter image description here

EDIT:

You can set a threshold and check whether max_val is greater than this threshold. If it is greater then proceed to draw the matching region.

threshold = 0.85
min_val, max_val, min_loc, max_loc = cv2.minMaxLoc(results)
if max_val > threshold:

    top_left = max_loc
    bottom_right = (top_left[0] + w, top_left[1] + h)
    cv2.rectangle(image.copy(), top_left, bottom_right, 255, 2)

Upvotes: 4

Related Questions