opencv学习之路（24）、轮廓查找与绘制（三）——凸包

一、简介

二、绘制点集的凸包

#include<opencv2/opencv.hpp>
using namespace cv;

void main()
{
    //---绘制点集的凸包
    Mat img(400, 400, CV_8UC3, Scalar::all(0));  //定义绘制图像
    RNG rng;  //定义随机数对象
    while(1)
    {
        char key;
        int count = (unsigned int)rng % 100;  //定义点的个数  
        vector<Point> points;  //定义点集
        for(int i=0; i<count; i++)
        {
            Point pt;
            pt.x = rng.uniform(img.cols/4, img.cols*3/4);  //设定点的x范围
            pt.y = rng.uniform(img.rows/4, img.rows*3/4);  //设定点的y范围
            points.push_back(pt);
        }

        //检测凸包
        vector<int> hull;
        convexHull(Mat(points), hull, true);

        img = Scalar::all(0);
        for(int i = 0; i < count; i++ )
            circle(img, points[i], 3, Scalar(rng.uniform(0, 255), rng.uniform(0, 255), rng.uniform(0, 255)), CV_FILLED, CV_AA);

        //准备参数
        int hullcount = (int)hull.size(); //凸包的边数
        Point point0 = points[hull[hullcount-1]]; //连接凸包边的坐标点

        //绘制凸包的边
        for(int  i = 0; i < hullcount; i++ )
        {
            Point point = points[hull[i]];
            circle(img, point, 8, Scalar(0, 255, 0), 2, 8);
            line(img, point0, point, Scalar(255, 255, 255), 2, CV_AA);
            point0 = point;
        }

        //显示效果图
        imshow("img", img);

        //按下ESC,Q,或者q，程序退出
        key = (char)waitKey();
        if( key == 27 || key == 'q' || key == 'Q' ) 
            break;
    }
51}

三、绘制轮廓的凸包

#include<opencv2/opencv.hpp>
using namespace cv;

void main()
{
    Mat srcImg = imread("E://12.jpg");
    imshow("src", srcImg);
    Mat dstImg2 = srcImg.clone();
    Mat tempImg(srcImg.rows, srcImg.cols, CV_8UC3, Scalar::all(0));  //用于绘制凸包
    Mat dstImg(srcImg.rows, srcImg.cols, CV_8UC3, Scalar::all(0));  //用于绘制轮廓
    cvtColor(srcImg, srcImg, CV_BGR2GRAY);
    threshold(srcImg, srcImg, 100, 255, CV_THRESH_BINARY); //二值化
    
    vector<vector<Point>> contours;
    vector<Vec4i> hierarcy;
    findContours(srcImg, contours, hierarcy, CV_RETR_EXTERNAL, CV_CHAIN_APPROX_NONE);
    vector<vector<Point>> hull(contours.size());
    for(int i=0; i<contours.size(); i++)
    {
        convexHull(Mat(contours[i]), hull[i], true);     //查找凸包
        drawContours(dstImg, contours, i, Scalar(255, 255, 255), -1, 8);  //绘制轮廓
        //drawContours(dstImg, hull, i, Scalar(rand()%255, rand()%255, rand()%255), 2, 8);
        drawContours(tempImg, hull, i, Scalar(255, 255, 255), -1, 8);
    }
    imshow("hull", tempImg);
    imshow("contours", dstImg);

    Mat diffImg;
    absdiff(tempImg, dstImg, diffImg);  //图像相减
    Mat element = getStructuringElement(MORPH_RECT, Size(3, 3), Point(-1, -1));
    erode(diffImg, diffImg, element);
    imshow("diff", diffImg);

    vector<vector<Point>> contours2;
    vector<Vec4i> hierarcy2;
    cvtColor(diffImg, diffImg, CV_BGR2GRAY); //转为灰度图
    threshold(diffImg, diffImg, 100, 255, CV_THRESH_BINARY); //二值化
    findContours(diffImg, contours2, hierarcy2, CV_RETR_EXTERNAL, CV_CHAIN_APPROX_NONE);
    drawContours(dstImg2, contours2, -1, Scalar(0, 0, 255), 2, 8);  //红色绘制缺陷轮廓
    imshow("defects", dstImg2);
    waitKey(0);
}

 原文：https://www.cnblogs.com/little-monkey/p/7424951.html

---