This code cuts an area with the largest symbol from image, that you offer.

CompVision.py

import cv2 as cv
import numpy as np
from matplotlib import pyplot as plt
from PIL import Image

def take_index(a,b,c,d):
    areas = [(a[i] * b[i], i) for i in range(len(a))]
    return sorted(areas)[-1][1]
img = cv.imread('color.png')

imgray = np.average(img,axis=2)

ret,thresh = cv.threshold(imgray,100,255,cv.THRESH_BINARY_INV)

#thresh.reshape((imgray.shape[0],imgray.shape[1],1))*np.array([1.,1.,1.]))


thresh = thresh.astype(np.uint8)


im2,contours, hierarchy = cv.findContours(thresh,cv.RETR_LIST,cv.CHAIN_APPROX_SIMPLE)
cv.drawContours(img,contours,-1,(0,255,0),3)
X = []
Y = []
W = []
H = []
plt.imshow(img)
plt.show()
for c in contours:
      x,y,w,h = cv.boundingRect(c)
      rect = cv.rectangle(img, (x, y), (x+w, y+h), (255,255,255), 2)
      cv.fillPoly(img,(np.array([rect],dtype=np.int32)),color = (255,255,255))
      X.append(x)
      Y.append(y)
      W.append(w)
      H.append(h)
#plt.imshow(img)
x_c = X[take_index(W,H,X,Y)]
y_c = Y[take_index(W,H,X,Y)]
w_c = W[take_index(W,H,X,Y)]
h_c = H[take_index(W,H,X,Y)]
cut_image = img[y_c:y_c+h_c, x_c:x_c+w_c]

#cv.drawContours(cut_image,contours,(0,0,255))
print(take_index(W,H,X,Y))
plt.imshow(cut_image)
plt.show()
print("-------------------------------------------------------------------------------------------------------------------------------------------------")
print(img.shape)
print("-------------------------------------------------------------------------------------------------------------------------------------------------")
#print(w*h)
plt.show()