OpenCV 闭合轮廓检测

OpenCV 闭合轮廓检测

这个好像是骨头什么的，但是要求轮廓闭合，于是对图片进行一下膨胀操作，再次检测轮廓就好了。

// A closed contour.cpp : 定义控制台应用程序的入口点。
//

#include "stdafx.h"

// FindRotation-angle.cpp : 定义控制台应用程序的入口点。
//

// findContours.cpp : 定义控制台应用程序的入口点。
//

#include "stdafx.h"

 

#include <iostream>
#include <vector>
#include <opencv2/opencv.hpp>
#include <opencv2/core/core.hpp>
#include <opencv2/imgproc/imgproc.hpp>
#include <opencv2/highgui/highgui.hpp>
//#include "highlight"
//#include "highgui.h"

#pragma comment(lib,"opencv_core2410d.lib")       
#pragma comment(lib,"opencv_highgui2410d.lib")       
#pragma comment(lib,"opencv_imgproc2410d.lib") 

#define PI 3.1415926

using namespace std;
using namespace cv;

int main()
{
 // Read input binary image

 char *image_name = "test.bmp";
 cv::Mat image = cv::imread(image_name);
 if (!image.data)
  return 0;

 

 
 // 从文件中加载原图 
  // IplImage *pSrcImage = cvLoadImage(image_name, CV_LOAD_IMAGE_UNCHANGED); 
  Mat gray(image.size(),CV_8U);
   
  cvtColor(image,gray,CV_BGR2GRAY);
  // 转为2值图
  threshold(gray,gray,145,255,cv::THRESH_BINARY_INV);
 //cvThreshold(pSrcImage,pSrcImage,145,255,cv::THRESH_BINARY_INV);
   
 
    //image = gray;

    cv::namedWindow("Binary Image");
    cv::imshow("Binary Image",gray);

 

    cv::Mat element(5,5,CV_8U,cv::Scalar(255));

    cv::dilate(gray,gray,element);
    //cv::erode(image,image,element);

    cv::namedWindow("dilate Image");
    cv::imshow("dilate Image",gray);

 // Get the contours of the connected components
 std::vector<std::vector<cv::Point>> contours;

 cv::findContours(gray,
  contours, // a vector of contours
  CV_RETR_EXTERNAL , // retrieve the external contours
  CV_CHAIN_APPROX_NONE); // retrieve all pixels of each contours

 // Print contours' length
 std::cout << "Contours: " << contours.size() << std::endl;
 std::vector<std::vector<cv::Point>>::const_iterator itContours= contours.begin();
 for ( ; itContours!=contours.end(); ++itContours)
 {

  std::cout << "Size: " << itContours->size() << std::endl;
 }

 // draw black contours on white image
 cv::Mat result(image.size(),CV_8U,cv::Scalar(255));
 cv::drawContours(result,contours,
  -1, // draw all contours
  cv::Scalar(0), // in black
  2); // with a thickness of 2

 cv::namedWindow("Contours");
 cv::imshow("Contours",result);

 

 

 

 

 // Eliminate too short or too long contours

 /*
 int cmin= 100;  // minimum contour length
 int cmax= 1000; // maximum contour length
 std::vector<std::vector<cv::Point>>::const_iterator itc= contours.begin();
 while (itc!=contours.end()) {

  if (itc->size() < cmin || itc->size() > cmax)
   itc= contours.erase(itc);
  else
   ++itc;
 }
 
 */

 // draw contours on the original image
 cv::Mat original= cv::imread(image_name);
 cv::drawContours(original,contours,
  -1, // draw all contours
  cv::Scalar(255,255,0), // in white
  2); // with a thickness of 2

 cv::namedWindow("Contours on Animals");
 cv::imshow("Contours on Animals",original);

 

 // Let's now draw black contours on white image
 result.setTo(cv::Scalar(255));
 cv::drawContours(result,contours,
  -1, // draw all contours
  cv::Scalar(0), // in black
  1); // with a thickness of 1
 image= cv::imread("binary.bmp",0);

 // testing the bounding box
 

 

 std::vector<std::vector<cv::Point>>::const_iterator itc_rec= contours.begin();
 while (itc_rec!=contours.end())
 {
  cv::Rect r0= cv::boundingRect(cv::Mat(*(itc_rec)));
  cv::rectangle(result,r0,cv::Scalar(0),2);
   ++itc_rec;
 }

 /*
 // testing the enclosing circle
 float radius;
 cv::Point2f center;
 cv::minEnclosingCircle(cv::Mat(contours[1]),center,radius);
 cv::circle(result,cv::Point(center),static_cast<int>(radius),cv::Scalar(0),2);

 // cv::RotatedRect rrect= cv::fitEllipse(cv::Mat(contours[1]));
 // cv::ellipse(result,rrect,cv::Scalar(0),2);

 // testing the approximate polygon
 std::vector<cv::Point> poly;
 cv::approxPolyDP(cv::Mat(contours[2]),poly,5,true);

 std::cout << "Polygon size: " << poly.size() << std::endl;

 // Iterate over each segment and draw it
 std::vector<cv::Point>::const_iterator itp= poly.begin();
 while (itp!=(poly.end()-1)) {
  cv::line(result,*itp,*(itp+1),cv::Scalar(0),2);
  ++itp;
 }
 // last point linked to first point
 cv::line(result,*(poly.begin()),*(poly.end()-1),cv::Scalar(20),2);

 // testing the convex hull
 std::vector<cv::Point> hull;
 cv::convexHull(cv::Mat(contours[3]),hull);

 // Iterate over each segment and draw it
 std::vector<cv::Point>::const_iterator it= hull.begin();
 while (it!=(hull.end()-1)) {
  cv::line(result,*it,*(it+1),cv::Scalar(0),2);
  ++it;
 }
 // last point linked to first point
 cv::line(result,*(hull.begin()),*(hull.end()-1),cv::Scalar(20),2);

 // testing the moments

 // iterate over all contours
 itc= contours.begin();
 while (itc!=contours.end()) {

  // compute all moments
  cv::Moments mom= cv::moments(cv::Mat(*itc++));

  // draw mass center
  cv::circle(result,
   // position of mass center converted to integer
   cv::Point(mom.m10/mom.m00,mom.m01/mom.m00),
   2,cv::Scalar(0),2); // draw black dot
 }

 */

 cv::namedWindow("Some Shape descriptors");
 cv::imshow("Some Shape descriptors",result);

 cv::waitKey();
 return 0;

}

实现效果：

--------------------------------------分割线 --------------------------------------

Ubuntu Linux下安装OpenCV2.4.1所需包

Ubuntu 12.04 安装 OpenCV2.4.2

CentOS下OpenCV无法读取视频文件

Ubuntu 12.04下安装OpenCV 2.4.5总结

Ubuntu 10.04中安装OpenCv2.1九步曲

基于QT和OpenCV的人脸识别系统

[翻译]Ubuntu 14.04, 13.10 下安装 OpenCV 2.4.9 

--------------------------------------分割线 --------------------------------------

OpenCV的详细介绍：请点这里
OpenCV的下载地址：请点这里

本文永久更新链接地址：