python-opencv最大内切圆查找

用pointPolygonTest查找内切圆在大图像上效率太低，先查找缩放后图像的内切圆，再去原图上查找可以大幅降低算法时间

import cv2
import numpy as np
import time

t0 = time.time()
mask_gray = cv2.imread("E:/20220519154514.bmp",0)
rows,cols = mask_gray.shape
# Get the contours
print(f'原图尺寸{mask_gray.shape}')
# 识别轮廓
_,mask_grayqqqq = cv2.threshold(mask_gray,50,255,cv2.THRESH_BINARY)
contours2, _ = cv2.findContours(mask_grayqqqq, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)

mm = cv2.resize(mask_grayqqqq,(int(cols/8/4),int(rows/8/4)))
contours, _ = cv2.findContours(mm, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)

for i in contours:
    if 100 < cv2.contourArea(i):
        # print(cv2.contourArea(i))
        cont = i

for i in contours2:
    if 5000 < cv2.contourArea(i):
        # print(cv2.contourArea(i))
        cont2 = i

# 计算到轮廓的距离
x, y, w, h = cv2.boundingRect(cont)
raw_dist = np.zeros_like(mm, dtype=np.float32)

raw_dist2 = np.zeros_like(mask_grayqqqq, dtype=np.float32)

for i in range(int(x+w/4),int(x+w/4*3)):
    for j in range(int(y+h/4),int(y+h/4*3)):
        raw_dist[i, j] = cv2.pointPolygonTest(cont, (j, i), True)

# 获取最大值即内接圆半径，中心点坐标
minVal, maxVal, _, maxDistPt = cv2.minMaxLoc(raw_dist)
print(f'缩放后的圆心{maxDistPt}')

for i in range(maxDistPt[1]*8*4-25,maxDistPt[1]*8*4+25):
    for j in range(maxDistPt[0]*8*4-25,maxDistPt[0]*8*4+25):
        raw_dist2[i, j] = cv2.pointPolygonTest(cont2, (j, i), True)

minVal, maxVal, _, maxDistPt = cv2.minMaxLoc(raw_dist2)
maxVal = abs(maxVal)
# 画出最大内接圆
result = cv2.cvtColor(mask_gray, cv2.COLOR_GRAY2BGR)
print(f'原图的圆心{maxDistPt}')

center_of_circle = maxDistPt
cv2.circle(result, maxDistPt, int(maxVal), (0, 255, 0), 2, 1, 0)
# cv2.imwrite('./sawww.bmp',result)
result = cv2.resize(result,(1024,1024))

print(f'耗时{time.time() - t0}')
cv2.imshow('Maximum inscribed circle', result)
cv2.waitKey(0)

原图尺寸(5120, 5120)
缩放后的圆心(93, 85)
原图的圆心(3000, 2714)
耗时0.19447755813598633

原图尺寸(1024, 1280)
缩放后的圆心(72, 68)
原图的圆心(582, 547)
耗时0.022937297821044922