基于Opencv的MeanShift跟踪算法实现

#include "cv.h"

#include "highgui.h"
#include <stdio.h>
#include <ctype.h>

IplImage *image = 0, *hsv = 0, *hue = 0, *mask = 0, *backproject = 0, *histimg = 0;//用HSV中的Hue分量进行跟踪
CvHistogram *hist = 0;//直方图类

int backproject_mode = 0;
int select_object = 0;
int track_object = 0;
int show_hist = 1;
CvPoint origin;
CvRect selection;
CvRect track_window;
CvBox2D track_box; // Meanshift跟踪算法返回的Box类
CvConnectedComp track_comp;
int hdims = 50; // 划分直方图bins的个数,越多越精确

float hranges_arr[] = {0,180};//像素值的范围
float* hranges = hranges_arr;//用于初始化CvHistogram类
int vmin = 10, vmax = 256, smin = 30;

void on_mouse( int event, int x, int y, int flags,void *NotUsed)//该函数用于选择跟踪目标
{
      if( !image )
         return;

     if( image->origin )
         y = image->height - y;

     if( select_object )//如果处于选择跟踪物体阶段,则对selection用当前的鼠标位置进行设置
     {
         selection.x = MIN(x,origin.x);
         selection.y = MIN(y,origin.y);
         selection.width = selection.x + CV_IABS(x - origin.x);
         selection.height = selection.y + CV_IABS(y - origin.y);

         selection.x = MAX( selection.x, 0 );
         selection.y = MAX( selection.y, 0 );
         selection.width = MIN( selection.width, image->width );
         selection.height = MIN( selection.height, image->height );
         selection.width -= selection.x;
         selection.height -= selection.y;

      }

      switch( event )
      {
            case CV_EVENT_LBUTTONDOWN://开始点击选择跟踪物体
                   origin = cvPoint(x,y);
                   selection = cvRect(x,y,0,0);//坐标
                   select_object = 1;//表明开始进行选取
            break;
           case CV_EVENT_LBUTTONUP:
                   select_object = 0;//选取完成
                   if( selection.width > 0 && selection.height > 0 )
                       track_object = -1;//如果选择物体有效,则打开跟踪功能

           break;
      }
}


CvScalar hsv2rgb( float hue )//用于将Hue量转换成RGB量
{
    int rgb[3], p, sector;
    static const int sector_data[][3]={{0,2,1}, {1,2,0}, {1,0,2}, {2,0,1}, {2,1,0}, {0,1,2}};
    hue *= 0.033333333333333333333333333333333f;
    sector = cvFloor(hue);
    p = cvRound(255*(hue - sector));
    p ^= sector & 1 ? 255 : 0;

    rgb[sector_data[sector][0]] = 255;
    rgb[sector_data[sector][1]] = 0;
    rgb[sector_data[sector][2]] = p;

    return cvScalar(rgb[2], rgb[1], rgb[0],0);//返回对应的颜色值
}

int main( int argc, char** argv )
{
    CvCapture* capture = 0;
    IplImage* frame = 0;

    if( argc == 1 || (argc == 2 && strlen(argv[1]) == 1 && isdigit(argv[1][0])))
         capture = cvCaptureFromCAM( argc == 2 ? argv[1][0] - '0' : 0 );//打开摄像头
    else if( argc == 2 )
         capture = cvCaptureFromAVI( argv[1] );//打开AVI文件

    if( !capture )
    {
        fprintf(stderr,"Could not initialize capturing.../n");//打开视频流失败处理
        return -1;
     }

    printf( "Hot keys: /n/tESC - quit the program/n/tc - stop the tracking/n/tb - switch to/from backprojection view/n/th - show/hide object histogram/nTo initialize tracking, select the object with mouse/n" );//打印出程序功能列表
    cvNamedWindow( "CamShiftDemo", 1 );//建立视频窗口
    cvSetMouseCallback( "CamShiftDemo", on_mouse ); // 设置鼠标回调函数

    cvCreateTrackbar( "Vmin", "CamShiftDemo", &vmin, 256, 0 );//建立滑动条
    cvCreateTrackbar( "Vmax", "CamShiftDemo", &vmax, 256, 0 );
    cvCreateTrackbar( "Smin", "CamShiftDemo", &smin, 256, 0 );

    for(;;)//进入视频帧处理主循环
   {
       int i, bin_w, c;
       frame = cvQueryFrame( capture );
       if( !frame )
               break;

      if( !image )//刚开始先建立一些缓冲区
      {

          image = cvCreateImage( cvGetSize(frame), 8, 3 );//
          image->origin = frame->origin;
          hsv = cvCreateImage( cvGetSize(frame), 8, 3 );
          hue = cvCreateImage( cvGetSize(frame), 8, 1 );
          mask = cvCreateImage( cvGetSize(frame), 8, 1 );//分配掩膜图像空间
          backproject = cvCreateImage( cvGetSize(frame), 8, 1 );//分配反向投影图空间,大小一样,单通道
          hist = cvCreateHist( 1, &hdims, CV_HIST_ARRAY, &hranges, 1 ); //分配建立直方图空间

          histimg = cvCreateImage( cvSize(320,200), 8, 3 );//分配用于画直方图的空间
         cvZero( histimg );//背景为黑色
       }

       cvCopy( frame, image, 0 );
       cvCvtColor( image, hsv, CV_BGR2HSV ); // 把图像从RGB表色系转为HSV表色系

       if( track_object )//   如果当前有需要跟踪的物体   

      {
             int _vmin = vmin, _vmax = vmax;

             cvInRangeS( hsv, cvScalar(0,smin,MIN(_vmin,_vmax),0),cvScalar(180,256,MAX(_vmin,_vmax),0), mask ); //制作掩膜板,只处理像素值为H:0~180,S:smin~256,V:vmin~vmax之间的部分
             cvSplit( hsv, hue, 0, 0, 0 ); // 取得H分量

             if( track_object < 0 )//如果需要跟踪的物体还没有进行属性提取,则进行选取框类的图像属性提取
             {
                    float max_val = 0.f;
                   cvSetImageROI( hue, selection ); // 设置原选择框
                   cvSetImageROI( mask, selection ); // 设置Mask的选择框

                   cvCalcHist( &hue, hist, 0, mask ); // 得到选择框内且满足掩膜板内的直方图

                   cvGetMinMaxHistValue( hist, 0, &max_val, 0, 0 ); 
                   cvConvertScale( hist->bins, hist->bins, max_val ? 255. / max_val : 0., 0 ); // 对直方图转为0~255
                   cvResetImageROI( hue ); // remove ROI
                   cvResetImageROI( mask );
                   track_window = selection;
                   track_object = 1;

                  cvZero( histimg );
                  bin_w = histimg->width / hdims;

                  for( i = 0; i < hdims; i++ )
                  {
                      int val = cvRound( 
                      cvGetReal1D(hist->bins,i)*histimg->height/255 );
                      CvScalar color = hsv2rgb(i*180.f/hdims);
                      cvRectangle( histimg, cvPoint(i*bin_w,histimg->height),
                      cvPoint((i+1)*bin_w,histimg->height - val),color, -1, 8, 0 );//画直方图到图像空间
                  }
        }

         cvCalcBackProject( &hue, backproject, hist ); // 得到hue的反向投影图

         cvAnd( backproject, mask, backproject, 0 );得到反向投影图mask内的内容
         cvCamShift( backproject, track_window,cvTermCriteria( CV_TERMCRIT_EPS | CV_TERMCRIT_ITER, 10, 1 ),&track_comp, &track_box );//使用MeanShift算法对backproject中的内容进行搜索,返回跟踪结果
         track_window = track_comp.rect;//得到跟踪结果的矩形框

         if( backproject_mode )
                cvCvtColor( backproject, image, CV_GRAY2BGR ); // 显示模式
         if( image->origin )
         track_box.angle = -track_box.angle;
        cvEllipseBox( image, track_box, CV_RGB(255,0,0), 3, CV_AA, 0 );//画出跟踪结果的位置
   }

  if( select_object && selection.width > 0 && selection.height > 0 )//如果正处于物体选择,画出选择框
  {
       cvSetImageROI( image, selection );
       cvXorS( image, cvScalarAll(255), image, 0 );
      cvResetImageROI( image );
  }

  cvShowImage( "CamShiftDemo", image );//显示视频和直方图
  cvShowImage( "Histogram", histimg );

  c = cvWaitKey(10);
  if( c == 27 )
      break;

  switch( c )
  {
     case 'b':
           backproject_mode ^= 1;
           break;
     case 'c':
           track_object = 0;
           cvZero( histimg );
           break;
    case 'h':
          show_hist ^= 1;
          if( !show_hist )
                cvDestroyWindow( "Histogram" );
         else
               cvNamedWindow( "Histogram", 1 );
         break;
    default:
           ;
      }
   }

   cvReleaseCapture( &capture );
   cvDestroyWindow("CamShiftDemo");

    return 0;
}

你可能感兴趣的:(基于Opencv的MeanShift跟踪算法实现)