opencv学习之路（27）、轮廓查找与绘制（六）——外接圆、椭圆拟合、逼近多边形曲线、计算轮廓面积及长度、提取不规则轮廓

一、最小外接圆

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

void main() {
    Mat srcImg = imread("E://10.png");
    imshow("src", srcImg);
    Mat dstImg = srcImg.clone();
    GaussianBlur(srcImg, srcImg, Size(3, 3), 0, 0);
    cvtColor(srcImg, srcImg, CV_BGR2GRAY);
    Canny(srcImg, srcImg, 100, 200);
    imshow("Canny", srcImg);

    //查找轮廓
    vector<vector<Point>> contours;
    vector<Vec4i> hierarcy;
    findContours(srcImg, contours, hierarcy, CV_RETR_TREE, CV_CHAIN_APPROX_NONE);
    cout << "num=" << contours.size() << endl;

    Point2f center;  //定义圆中心坐标
    float radius;  //定义圆半径
    for (int i = 0; i<contours.size(); i++)  //依次遍历每个轮廓
    {
        minEnclosingCircle(Mat(contours[i]), center, radius);
        drawContours(dstImg, contours, i, Scalar(0, 0, 255), 2, 8);
        circle(dstImg, center, radius, Scalar(0, 255, 0), 2, 8);  //绘制第i个轮廓的最小外接圆
    }
    imshow("dst", dstImg);

    waitKey(0);
}

 二、椭圆拟合

#include "opencv2/opencv.hpp"
using namespace cv;

void main() {
    Mat srcImg = imread("E://10.png");
    imshow("src", srcImg);
    Mat dstImg = srcImg.clone();
    GaussianBlur(srcImg, srcImg, Size(3, 3), 0, 0);
    cvtColor(srcImg, srcImg, CV_BGR2GRAY);
    Canny(srcImg, srcImg, 100, 200);
    imshow("Canny", srcImg);

    //查找轮廓
    vector<vector<Point>> contours;
    vector<Vec4i> hierarcy;
    findContours(srcImg, contours, hierarcy, CV_RETR_TREE, CV_CHAIN_APPROX_NONE);

    vector<RotatedRect> box(contours.size());
    Point2f rect[4];
    for (int i = 0; i<contours.size(); i++)
    {
        box[i] = fitEllipse(Mat(contours[i]));
        //ellipse(dstImg, box[i].center, Size(box[i].size.width/2, box[i].size.height/2), box[i].angle, 0, 360, Scalar(0, 255, 0), 2, 8);
        ellipse(dstImg, box[i], Scalar(0, 255, 0), 2, 8);
    }
    imshow("dst", dstImg);

    waitKey(0);
}

三、逼近多边形曲线

#include "opencv2/opencv.hpp"
using namespace cv;

void main() {
    Mat srcImg = imread("E://02.jpg");
    imshow("src", srcImg);
    Mat dstImg = srcImg.clone();
    Mat dstImg2(srcImg.size(), CV_8UC3, Scalar::all(0));//全黑图像

    GaussianBlur(srcImg, srcImg, Size(3, 3), 0, 0);
    cvtColor(srcImg, srcImg, CV_BGR2GRAY);
    //Canny(srcImg, srcImg, 100, 200);
    threshold(srcImg, srcImg, 200, 255, CV_THRESH_BINARY_INV);
    imshow("threshold", srcImg);

    vector<vector<Point>> contours;
    vector<Vec4i> hierarcy;
    findContours(srcImg, contours, hierarcy, CV_RETR_TREE, CV_CHAIN_APPROX_NONE);
    vector<vector<Point>> contours_poly(contours.size());

    for (int i = 0; i<contours.size(); i++)
    {
        approxPolyDP(Mat(contours[i]), contours_poly[i],15, true);//true曲线封闭，反之不封闭
        drawContours(dstImg, contours, i, Scalar(0, 255, 0), 2, 8);//绘制轮廓
        drawContours(dstImg2, contours_poly, i, Scalar(0, 255, 255), 2, 8);  //绘制多边形逼近
    }
    imshow("dst", dstImg);
    imshow("approx", dstImg2);

    waitKey(0);
}

 四、计算轮廓面积及长度（可用于轮廓筛选）

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

void main() {
    Mat srcImg = imread("E://33.jpg");
    imshow("src", srcImg);
    Mat dstImg = srcImg.clone();
    Mat dstImg2(srcImg.size(), CV_8UC3, Scalar::all(0));

    GaussianBlur(srcImg, srcImg, Size(3, 3), 0, 0);
    cvtColor(srcImg, srcImg, CV_BGR2GRAY);
    //Canny(srcImg, srcImg, 100, 200);
    threshold(srcImg, srcImg, 200, 255, CV_THRESH_BINARY);
    imshow("threshold", srcImg);

    vector<vector<Point>> contours;
    vector<Vec4i> hierarcy;
    findContours(srcImg, contours, hierarcy, CV_RETR_TREE, CV_CHAIN_APPROX_NONE);
    cout << "num=" << contours.size() << endl;

    for (int i = 0; i<contours.size(); i++)
    {
        double area = contourArea(contours[i]);//计算第i个轮廓的面积
        cout<<"area--"<<i<<"---"<<area<<endl;

        double length = arcLength(contours[i], true);
        cout << "length--" << i << "---" << length << endl;
        //if(area>10000)  //面积大约1W
        //if(area> 100 && area<300)

        if (length<300 && area>300)  
            drawContours(dstImg, contours, i, Scalar(0, 0, 255), 2, 8);
    }
    imshow("dst", dstImg);
    waitKey(0);
}

五、提取不规则轮廓

#include "opencv2/opencv.hpp"
using namespace cv;

void main() {
    Mat srcImg = imread("E://220.jpg");
    imshow("src", srcImg);
    Mat dstImg = srcImg.clone();  //原图备份
    Mat tempImg = srcImg.clone();  //原图备份
    Mat tempImg2(srcImg.size(), CV_8UC3, Scalar::all(0));  //定义全黑的和原图一样大小的图像
    Mat draw(srcImg.size(), CV_8UC3, Scalar::all(0)); 
    Mat tempImg3(srcImg.size(), CV_8UC3, Scalar::all(0)); 

    GaussianBlur(srcImg, srcImg, Size(3, 3), 0, 0);
    cvtColor(srcImg, srcImg, CV_BGR2GRAY);
    threshold(srcImg, srcImg, 100, 255, CV_THRESH_BINARY); //二值化
    imshow("threshold", srcImg);

    vector<vector<Point>> contours;
    vector<Vec4i> hierarcy;
    findContours(srcImg, contours, hierarcy, CV_RETR_TREE, CV_CHAIN_APPROX_NONE);
    while (1)
    {
        for (int i = 0; i<contours.size(); i++)
        {
            tempImg2.copyTo(draw);  //每次进入将draw清空为全黑
            tempImg2.copyTo(tempImg3); 
            //drawContours(dstImg, contours, i, Scalar(0, 255, 0), 5, 8);    
            drawContours(draw, contours, i, Scalar(255, 255, 255), -1, 8);
            Mat mask;  //定义掩码
            cvtColor(draw, mask, CV_BGR2GRAY);
            tempImg.copyTo(tempImg3, mask);  //将tempImg 复制到tempImg3(只有mask部分被复制)
            imshow("draw", draw);
            imshow("result", tempImg3);
            char key = waitKey();
            if (key == 27)  //按下Esc键跳出for循环
                break;
        }
        break;
    }
}