python-opencv最大内切圆查找

weixin_61011490

已于 2022-06-02 16:33:32 修改

阅读量1.6k

点赞数

文章标签： opencv python 计算机视觉

于 2022-06-02 16:24:46 首次发布

本文链接：https://blog.csdn.net/weixin_61011490/article/details/125102549

版权

本文探讨了使用OpenCV在Python中查找最大内切圆时遇到的效率问题。通过先在缩放后的图像上进行查找，然后再在原始图像上验证，可以显著提高算法的运行速度。

摘要由CSDN通过智能技术生成

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_li