我最近一直在做Python和C版本OpenCV中HoughCircle算法的计时测试，就是想验证直觉上C处理速度更快的猜想！先跟大家说下我的环境配置：

• Python 3.6.4
• GCC编译器版本：gcc (Ubuntu 5.4.0-6ubuntu1~16.04.9) 5.4.0 20160609
• CMake 3.5.1
• OpenCV 3.4.1（Python版是通过Anaconda安装的，意外的是C++版也能正常跑起来）

注：原文中提到的测试图片未提供，默认使用适合圆形检测的常规图像（比如带圆形物体的场景图）

Python测试代码 #

我整理了完整可运行的Python测试脚本，包含计时逻辑：

import cv2
import time
import sys

def hough_transform(src, dp, minDist, param1=100, param2=100, minRadius=0, maxRadius=0):
    gray = cv2.cvtColor(src, cv2.COLOR_BGR2GRAY)
    circles = cv2.HoughCircles(gray, cv2.HOUGH_GRADIENT, dp, minDist,
                               param1=param1, param2=param2,
                               minRadius=minRadius, maxRadius=maxRadius)
    return circles

if __name__ == "__main__":
    if len(sys.argv) != 2:
        print("Usage: python hough_circle_test.py <image_path>")
        sys.exit(1)
    img_path = sys.argv[1]
    img = cv2.imread(img_path)
    if img is None:
        print("Could not read the image")
        sys.exit(1)
    
    # 计时测试
    start_time = time.time()
    circles = hough_transform(img, dp=1, minDist=20)
    end_time = time.time()
    
    print(f"Python HoughCircle execution time: {end_time - start_time:.4f} seconds")
    if circles is not None:
        print(f"Detected {len(circles[0])} circles")

C++测试代码 #

对应的C++版本测试代码，同样包含高精度计时：

#include <opencv2/opencv.hpp>
#include <iostream>
#include <chrono>

using namespace cv;
using namespace std;
using namespace chrono;

vector<Vec3f> hough_transform(Mat src, double dp, double minDist, double param1=100, double param2=100, int minRadius=0, int maxRadius=0) {
    Mat gray;
    cvtColor(src, gray, COLOR_BGR2GRAY);
    vector<Vec3f> circles;
    HoughCircles(gray, circles, HOUGH_GRADIENT, dp, minDist, param1, param2, minRadius, maxRadius);
    return circles;
}

int main(int argc, char** argv) {
    if (argc != 2) {
        cout << "Usage: ./hough_circle_test <image_path>" << endl;
        return -1;
    }
    Mat img = imread(argv[1]);
    if (img.empty()) {
        cout << "Could not read the image" << endl;
        return -1;
    }
    
    // 高精度计时
    auto start_time = high_resolution_clock::now();
    vector<Vec3f> circles = hough_transform(img, 1, 20);
    auto end_time = high_resolution_clock::now();
    duration<double> elapsed = end_time - start_time;
    
    cout << "C++ HoughCircle execution time: " << elapsed.count() << " seconds" << endl;
    if (!circles.empty()) {
        cout << "Detected " << circles.size() << " circles" << endl;
    }
    return 0;
}

C++编译配置（CMakeLists.txt） #

为了方便编译C++代码，附上CMake配置文件：

cmake_minimum_required(VERSION 3.5)
project(hough_circle_test)

find_package(OpenCV 3.4 REQUIRED)

add_executable(hough_circle_test main.cpp)
target_link_libraries(hough_circle_test ${OpenCV_LIBS})

测试注意事项 #

为了让测试结果更准确、公平，建议：

• 确保Python和C++版本使用完全相同的HoughCircle参数（dp、minDist、param1等）
• 多次运行测试取平均值，避免单次运行的系统负载波动影响结果
• 尝试用不同尺寸、复杂度的图片测试，覆盖更多实际场景

内容的提问来源于stack exchange，提问作者Abhijit Balaji