How to find one image inside another using opencv?

Question

Python

There is such a code, it does a great job with large color images. I needed to find out if there is a gray arrow (2.png) on ​​another image (1.png) and there is not a single match. The arrow has been cropped from the original image. What could be the problem with the color "too gray" after processing?

spoiler

import cv2
import numpy as np
import matplotlib.pyplot as plt
import sys
import requests
import os

img_rgb = cv2.imread('1.png')
img_gray = cv2.cvtColor(img_rgb, cv2.COLOR_BGR2GRAY)
template = cv2.imread('2.png',0)
w, h = template.shape[::-1]

res = cv2.matchTemplate(img_gray,template,cv2.TM_SQDIFF_NORMED)
threshold = 0.4
loc = np.where( res >= threshold)

f = set()

for pt in zip(*loc[::-1]):
    cv2.rectangle(img_rgb, pt, (pt[0] + w, pt[1] + h), (0,0,255), 2)

    sensitivity = 60
    f.add((round(pt[0]/sensitivity), round(pt[1]/sensitivity)))

cv2.imwrite('result.png',img_rgb)

found_count = len(f)

if found_count >0:
    print ("found")
    
else:
    print ("not found")

Answer 1

Probably just need to swap 1.png and 2.png in the code.
templateis a larger image than the desired one.