高级图像处理初步（大米粒的处理）

#include <cv.h>
#include <highgui.h>
#include <math.h>
//#include <stdlib.h>
//#include <stdio.h>
 
 
int main(int argc, char* argv[])
{
   	 IplImage *src = 0; //定义源图像指针
  	 IplImage *tmp = 0; //定义临时图像指针
 	 IplImage *src_back = 0; //定义源图像背景指针
 	 IplImage *dst_gray = 0; //定义源文件去掉背景后的目标灰度图像指针
 	 IplImage *dst_bw = 0; //定义源文件去掉背景后的目标二值图像指针
 	 IplImage *dst_contours = 0; //定义轮廓图像指针
 	 IplConvKernel *element = 0; //定义形态学结构指针
 	 int Number_Object =0; 	    //定义目标对象数量
  	int contour_area_tmp = 0; //定义目标对象面积临时寄存器
  	int contour_area_sum = 0; //定义目标所有对象面积的和
  	int contour_area_ave = 0; //定义目标对象面积平均值
   	int contour_area_max = 0; //定义目标对象面积最大值
    	CvMemStorage *stor = 0;
   	 CvSeq * cont = 0;
    	CvContourScanner contour_scanner; 
   	 CvSeq * a_contour= 0;
 
   	 //1.读取和显示图像
 
   	 /* the first command line parameter must be image file name */
  	if ( argc == 2 && (src = cvLoadImage(argv[1], 0))!=0 )
    	{
       		 ;
   	}
   	else
  	{
       	 src = cvLoadImage("E:\\图片\\Rice.png", 0);
   	 }
    cvNamedWindow( "src", CV_WINDOW_AUTOSIZE );
    cvShowImage( "src", src );
    //cvSmooth(src, src, CV_MEDIAN, 3, 0, 0, 0); //中值滤波，消除小的噪声；
 
    //2.估计图像背景
    tmp = cvCreateImage( cvGetSize(src), src->depth, src->nChannels);
    src_back = cvCreateImage(cvGetSize(src), src->depth, src->nChannels);
    //创建结构元素
    element = cvCreateStructuringElementEx( 4, 4, 1, 1, CV_SHAPE_ELLIPSE, 0);
    //用该结构对源图象进行数学形态学的开操作后，估计背景亮度
    cvErode( src, tmp, element, 10);
    cvDilate( tmp, src_back, element, 10);
    cvNamedWindow( "src_back", CV_WINDOW_AUTOSIZE );
    cvShowImage( "src_back", src_back );
 
    //3.从源图象中减区背景图像
    dst_gray = cvCreateImage( cvGetSize(src), src->depth, src->nChannels);
    cvSub( src, src_back, dst_gray, 0);
    cvNamedWindow( "dst_gray", CV_WINDOW_AUTOSIZE );
    cvShowImage( "dst_gray", dst_gray );
 
    //4.使用阈值操作将图像转换为二值图像
    dst_bw = cvCreateImage( cvGetSize(src), src->depth, src->nChannels);
    cvThreshold( dst_gray, dst_bw ,50, 255, CV_THRESH_BINARY ); //取阈值为50把图像转为二值图像
 
    //cvAdaptiveThreshold( dst_gray, dst_bw, 255, CV_ADAPTIVE_THRESH_MEAN_C, CV_THRESH_BINARY, 3, 5 );
    cvNamedWindow( "dst_bw", CV_WINDOW_AUTOSIZE );
    cvShowImage( "dst_bw", dst_bw );
 
    //5.检查图像中的目标对象数量
	stor = cvCreateMemStorage(0);
     cont = cvCreateSeq(CV_SEQ_ELTYPE_POINT, sizeof(CvSeq), sizeof(CvPoint), stor);
     Number_Object = cvFindContours(dst_bw, stor, &cont, sizeof(CvContour), CV_RETR_TREE, CV_CHAIN_APPROX_SIMPLE, cvPoint(0,0) ); //找到所有轮廓
 
    printf("Number_Object: %d\n", Number_Object);
 
    //6.计算图像中对象的统计属性
    dst_contours = cvCreateImage( cvGetSize(src), src->depth, src->nChannels);
    cvThreshold( dst_contours, dst_contours ,0, 255, CV_THRESH_BINARY_INV); //在画轮廓前先把图像变成白色
 
    for(;cont;cont = cont->h_next)
    {
        cvDrawContours( dst_contours, cont, CV_RGB(255, 0, 0), CV_RGB(255, 0, 0), 0, 1, 8, cvPoint(0, 0) ); //绘制当前轮廓
 
        contour_area_tmp = fabs(cvContourArea( cont, CV_WHOLE_SEQ )); //获取当前轮廓面积
 
        if( contour_area_tmp > contour_area_max )
        {
            contour_area_max = contour_area_tmp; //找到面积最大的轮廓
 
        }
        contour_area_sum += contour_area_tmp; //求所有轮廓的面积和
    }
    contour_area_ave = contour_area_sum/ Number_Object; //求出所有轮廓的平均值
 
    printf("contour_area_ave: %d\n", contour_area_ave );
    printf("contour_area_max: %d\n", contour_area_max );
    cvNamedWindow( "dst_contours", CV_WINDOW_AUTOSIZE );
    cvShowImage( "dst_contours", dst_contours );
 
    cvWaitKey(-1); //等待退出
    cvReleaseImage(&src);
    cvReleaseImage(&tmp);
    cvReleaseImage(&src_back);
    cvReleaseImage(&dst_gray);
    cvReleaseImage(&dst_bw);
    cvReleaseImage(&dst_contours);
    cvReleaseMemStorage(&stor);
    cvDestroyWindow( "src" );
    cvDestroyWindow( "src_back" );
    cvDestroyWindow( "dst_gray" );
    cvDestroyWindow( "dst_bw" );
    cvDestroyWindow( "dst_contours" );
    //void cvDestroyAllWindows(void);
    return 0;
}

通过opencv2改写后的代码：

#include <opencv2/opencv.hpp>
#include <opencv2/imgproc/imgproc.hpp>
#include <iostream>
using namespace std;
using namespace cv;
int main()
{
	Mat src; //源图像
	Mat tmp; //临时图像
	Mat src_back; //源图像背景
	Mat dst_gray; //去掉背景后的目标灰度图像
	Mat dst_bw; //去掉背景后的目标二值图像
	Mat dst_contours;//轮廓图像
	src=imread("E:\\图片\\Rice.png",0);
	namedWindow("原图");
	imshow("原图",src);
	//medianBlur(src,src,7);//中值滤波
	tmp.create(src.size(),src.type());
	src_back.create(src.size(),src.type());
	//2.估计图像背景
	Mat element=getStructuringElement(MORPH_ELLIPSE,Size(4,4),Point(1,1));//用该结构对源图象进行数学形态	学的开操作后，估计背景亮度
	erode(src,tmp,element,Point(-1,-1),10);
	dilate(tmp,src_back,element,Point(-1,-1),10);
	namedWindow("src_back");
	imshow("src_back",src_back);
	//3.从源图象中减区背景图像
	dst_gray.create(src.size(),src.type());
	subtract(src, src_back,dst_gray);
	namedWindow("dst_gray");
	imshow("dst_gray",dst_gray);
	//4.使用阈值操作将图像转换为二值图像
	dst_bw.create(src.size(),src.type());
	threshold( dst_gray, dst_bw ,50, 255, CV_THRESH_BINARY);
	namedWindow("dst_bw");
	imshow("dst_bw",dst_bw);
	//5.检查图像中的目标对象数量
	vector<vector<Point>> contours;
	vector<Vec4i> hierarchy;
	findContours(dst_bw,contours,hierarchy,CV_RETR_TREE ,CV_CHAIN_APPROX_SIMPLE);
	int num=contours.size();
	cout<<"物体数量："<<num<<endl;
	//6.计算图像中对象的统计属性
	dst_contours.create(src.size(),src.type());
	threshold(dst_contours, dst_contours ,0, 255, CV_THRESH_BINARY_INV);//在画轮廓前先把图像变成白色
	namedWindow("dst_contours");
	imshow("dst_contours",dst_contours);
	size_t index=0;
	double contour_area_max=0;
	double contour_area_sum=0;
	double contour_area_tmp;
	for(size_t index=0;index<contours.size();index++)
	{
		drawContours(dst_contours,contours,-1,Scalar(255, 0, 0), 1, 8 );
		contour_area_tmp=contourArea(contours[index],false);
		if( contour_area_tmp > contour_area_max )
       		 {
            		contour_area_max = contour_area_tmp; //找到面积最大的轮廓
 
       		 }
       	 	contour_area_sum += contour_area_tmp; //求所有轮廓的面积和
 
	}
	double contour_area_ave = contour_area_sum/ num; //求出所有轮廓的平均值
	cout<<"contour_area_ave:"<<contour_area_ave<<endl;
	cout<<"contour_area_max:"<<contour_area_max<<endl;
	waitKey();
	return 0;
}