Intellipaat Back

Explore Courses Blog Tutorials Interview Questions
0 votes
2 views
in Python by (16.4k points)

I have a little python content utilizing OpenCV that does layout matching in a picture and pleasantly returns a bounding box, as beneath. The content is size invariant too which makes it more hearty. 

Given the bounding box returned, how might I supplant it with another format and save the adjusted picture?

Look at the Original template and main image

Furthermore, presently I might want to just replace the specific box with a resized form of template2 beneath, and save the new picture. How might I do that?

Here is my code:

# USAGE

# python match.py --template cod_logo.png --images images

# import the necessary packages

import numpy as np

import argparse

import imutils

import glob

import cv2

# construct the argument parser and parse the arguments

ap = argparse.ArgumentParser()

ap.add_argument("-t", "--template", required=True, help="Path to template image")

ap.add_argument("-i", "--images", required=True,

    help="Path to images dir where template will be matched")

ap.add_argument("-v", "--visualize",

    help="Flag 0 or 1 indicating whether or not to visualize each iteration")

args = vars(ap.parse_args())

# load the image image, convert it to grayscale, and detect edges

template = cv2.imread(args["template"])

template = cv2.cvtColor(template, cv2.COLOR_BGR2GRAY)

template = cv2.Canny(template, 50, 200)

(tH, tW) = template.shape[:2]

cv2.imshow("Template", template)

# loop over the images to find the template in

for imagePath in glob.glob(args["images"] + "/*.*"):

    # load the image, convert it to grayscale, and initialize the

    # bookkeeping variable to keep track of the matched region

    image = cv2.imread(imagePath)

    gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)

    found = None

    # loop over the scales of the image

    for scale in np.linspace(0.2, 1.0, 20)[::-1]:

        # resize the image according to the scale, and keep track

        # of the ratio of the resizing

        resized = imutils.resize(gray, width = int(gray.shape[1] * scale))

        r = gray.shape[1] / float(resized.shape[1])

        # if the resized image is smaller than the template, then break

        # from the loop

        if resized.shape[0] < tH or resized.shape[1] < tW:

            break

        # detect edges in the resized, grayscale image and apply template

        # matching to find the template in the image

        edged = cv2.Canny(resized, 50, 200)

        result = cv2.matchTemplate(edged, template, cv2.TM_CCOEFF)

        (_, maxVal, _, maxLoc) = cv2.minMaxLoc(result)

        # check to see if the iteration should be visualized

        if args.get("visualize", False):

            # draw a bounding box around the detected region

            clone = np.dstack([edged, edged, edged])

            cv2.rectangle(clone, (maxLoc[0], maxLoc[1]),

                (maxLoc[0] + tW, maxLoc[1] + tH), (0, 0, 255), 2)

            cv2.imshow("Visualize", clone)

            cv2.waitKey(0)

        # if we have found a new maximum correlation value, then ipdate

        # the bookkeeping variable

        if found is None or maxVal > found[0]:

            found = (maxVal, maxLoc, r)

    # unpack the bookkeeping varaible and compute the (x, y) coordinates

    # of the bounding box based on the resized ratio

    (_, maxLoc, r) = found

    (startX, startY) = (int(maxLoc[0] * r), int(maxLoc[1] * r))

    (endX, endY) = (int((maxLoc[0] + tW) * r), int((maxLoc[1] + tH) * r))

    # draw a bounding box around the detected result and display the image

    cv2.rectangle(image, (startX, startY), (endX, endY), (0, 0, 255), 2)

    cv2.imshow("Image", image)

    cv2.waitKey(0)

1 Answer

0 votes
by (26.4k points)

Here, I have also made few changes

1] I'm not using argument parser

2] tepmlate2 is COUNTER STRIKE.

3] Image2 is the image in which COUNTER STRIKE IS ON TOP OF COD.

STEPS : Extracting roi (region of interest) , then resizing the new image accordingly ... , then operlapping roi with new resized image, placing roi back on the image2.

Advantage => you can change the opacity of the roi and template by changing alpha and beta in addWeighted.

# USAGE

# python match.py --template cod_logo.png --images images

# import the necessary packages

import numpy as np

import argparse

import imutils

import glob

import cv2

#New template

template2 = cv2.imread("template2.png")

# construct the argument parser and parse the arguments

# ap = argparse.ArgumentParser()

# ap.add_argument("-t", "--template", required=True, help="Path to template image")

# ap.add_argument("-i", "--images", required=True,

#     help="Path to images dir where template will be matched")

# ap.add_argument("-v", "--visualize",

#     help="Flag 0 or 1 indicating whether or not to visualize each iteration")

# args = vars(ap.parse_args())

# load the image image, convert it to grayscale, and detect edges

template = cv2.imread("template.png")

template = cv2.cvtColor(template, cv2.COLOR_BGR2GRAY)

template = cv2.Canny(template, 50, 200)

(tH, tW) = template.shape\[:2\]

cv2.imshow("Template", template)

# loop over the images to find the template in

#for imagePath in glob.glob(args\["images"\] + "/*.*"):

    # load the image, convert it to grayscale, and initialize the

    # bookkeeping variable to keep track of the matched region

image = cv2.imread("mainImage.jpg")

gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)

found = None

# loop over the scales of the image

for scale in np.linspace(0.2, 1.0, 20)\[::-1\]:

    # resize the image according to the scale, and keep track

    # of the ratio of the resizing

    resized = imutils.resize(gray, width = int(gray.shape\[1\] * scale))

    r = gray.shape\[1\] / float(resized.shape\[1\])

    # if the resized image is smaller than the template, then break

    # from the loop

    if resized.shape\[0\] < tH or resized.shape\[1\] < tW:

        break

    # detect edges in the resized, grayscale image and apply template

    # matching to find the template in the image

    edged = cv2.Canny(resized, 50, 200)

    result = cv2.matchTemplate(edged, template, cv2.TM_CCOEFF)

    (_, maxVal, _, maxLoc) = cv2.minMaxLoc(result)

    # check to see if the iteration should be visualized

    '''

    if args.get("visualize", False):

        # draw a bounding box around the detected region

        clone = np.dstack(\[edged, edged, edged\])

        cv2.rectangle(clone, (maxLoc\[0\], maxLoc\[1\]),

            (maxLoc\[0\] + tW, maxLoc\[1\] + tH), (0, 0, 255), 2)

        cv2.imshow("Visualize", clone)

        cv2.waitKey(0)'''

    # if we have found a new maximum correlation value, then ipdate

    # the bookkeeping variable

    if found is None or maxVal > found\[0\]:

        found = (maxVal, maxLoc, r)

# unpack the bookkeeping varaible and compute the (x, y) coordinates

# of the bounding box based on the resized ratio

(_, maxLoc, r) = found

(startX, startY) = (int(maxLoc\[0\] * r), int(maxLoc\[1\] * r))

(endX, endY) = (int((maxLoc\[0\] + tW) * r), int((maxLoc\[1\] + tH) * r))

#MY CODE

image2 = image.copy()

resizedTemplate = cv2.resize(template2, (endX-startX, endY-startY), interpolation = cv2.INTER_AREA)

roi = image2[startY:endY, startX:endX]

img = cv2.addWeighted(resizedTemplate, 1, roi, 0, 0)

image2[startY:endY, startX:endX] = img

# draw a bounding box around the detected result and display the image

cv2.rectangle(image, (startX, startY), (endX, endY), (0, 0, 128), 2)

cv2.imshow("Image", image)

cv2.imshow("Image2", image2)

#cv2.imshow("resizedTemplate", resizedTemplate)

cv2.waitKey(0)

Interested to learn python in detail? Come and Join the python course.

Related questions

0 votes
1 answer
0 votes
1 answer
0 votes
1 answer
asked Mar 21, 2021 in Python by laddulakshana (16.4k points)

31k questions

32.8k answers

501 comments

693 users

Browse Categories

...