void cv::HoughLinesP( InputArray _image, OutputArray _lines,
      double rho, double theta, int threshold,
      double minLineLength, double maxGap )
{
 Ptr storage = cvCreateMemStorage(STORAGE_SIZE);
 Mat image = _image.getMat();
 CvMat c_image = image;
 CvSeq* seq = cvHoughLines2( &c_image, storage, CV_HOUGH_PROBABILISTIC,
     rho, theta, threshold, minLineLength, maxGap );
 seqToMat(seq, _lines);
}

static void
icvHoughLinesProbabilistic( CvMat* image,
       float rho, float theta, int threshold,
       int lineLength, int lineGap,
       CvSeq *lines, int linesMax )
{
 //accum為累加器矩陣，mask為掩碼矩陣
 cv::Mat accum, mask;
 cv::vector trigtab; //用于存儲(chǔ)事先計(jì)算好的正弦和余弦值
 //開辟一段內(nèi)存空間
 cv::MemStorage storage(cvCreateMemStorage(0));
 //用于存儲(chǔ)特征點(diǎn)坐標(biāo)，即邊緣像素的位置
 CvSeq* seq; 
 CvSeqWriter writer;
 int width, height; //圖像的寬和高
 int numangle, numrho; //角度和距離的離散數(shù)量
 float ang;
 int r, n, count;
 CvPoint pt;
 float irho = 1 / rho; //距離分辨率的倒數(shù)
 CvRNG rng = cvRNG(-1); //隨機(jī)數(shù)
 const float* ttab; //向量trigtab的地址指針
 uchar* mdata0; //矩陣mask的地址指針
 //確保輸入圖像的正確性
 CV_Assert( CV_IS_MAT(image) && CV_MAT_TYPE(image->type) == CV_8UC1 );
 
 width = image->cols; //提取出輸入圖像的寬
 height = image->rows; //提取出輸入圖像的高
 //由角度和距離分辨率，得到角度和距離的離散數(shù)量
 numangle = cvRound(CV_PI / theta);
 numrho = cvRound(((width + height) * 2 + 1) / rho);
 //創(chuàng)建累加器矩陣，即霍夫空間
 accum.create( numangle, numrho, CV_32SC1 );
 //創(chuàng)建掩碼矩陣，大小與輸入圖像相同
 mask.create( height, width, CV_8UC1 );
 //定義trigtab的大小，因?yàn)橐鎯?chǔ)正弦和余弦值，所以長度為角度離散數(shù)的2倍
 trigtab.resize(numangle*2);
 //累加器矩陣清零
 accum = cv::Scalar(0);
 //避免重復(fù)計(jì)算，事先計(jì)算好所需的所有正弦和余弦值
 for( ang = 0, n = 0; n < numangle; ang += theta, n++ )
 {
  trigtab[n*2] = (float)(cos(ang) * irho);
  trigtab[n*2+1] = (float)(sin(ang) * irho);
 }
 //賦值首地址
 ttab = &trigtab[0];
 mdata0 = mask.data;
 //開始寫入序列
 cvStartWriteSeq( CV_32SC2, sizeof(CvSeq), sizeof(CvPoint), storage, &writer );
 
 // stage 1. collect non-zero image points
 //收集圖像中的所有非零點(diǎn)，因?yàn)檩斎雸D像是邊緣圖像，所以非零點(diǎn)就是邊緣點(diǎn)
 for( pt.y = 0, count = 0; pt.y < height; pt.y++ )
 {
  //提取出輸入圖像和掩碼矩陣的每行地址指針
  const uchar* data = image->data.ptr + pt.y*image->step;
  uchar* mdata = mdata0 + pt.y*width;
  for( pt.x = 0; pt.x < width; pt.x++ )
  {
   if( data[pt.x] ) //是邊緣點(diǎn)
   {
    mdata[pt.x] = (uchar)1; //掩碼的相應(yīng)位置置1
    CV_WRITE_SEQ_ELEM( pt, writer ); 把該坐標(biāo)位置寫入序列
   }
   else //不是邊緣點(diǎn)
    mdata[pt.x] = 0; //掩碼的相應(yīng)位置清0
  }
 }
 //終止寫序列，seq為所有邊緣點(diǎn)坐標(biāo)位置的序列
 seq = cvEndWriteSeq( &writer );
 count = seq->total; //得到邊緣點(diǎn)的數(shù)量
 
 // stage 2. process all the points in random order
 //隨機(jī)處理所有的邊緣點(diǎn)
 for( ; count > 0; count-- )
 {
  // choose random point out of the remaining ones
  //步驟1，在剩下的邊緣點(diǎn)中隨機(jī)選擇一個(gè)點(diǎn)，idx為不大于count的隨機(jī)數(shù)
  int idx = cvRandInt(&rng) % count;
  //max_val為累加器的最大值，max_n為最大值所對(duì)應(yīng)的角度
  int max_val = threshold-1, max_n = 0;
  //由隨機(jī)數(shù)idx在序列中提取出所對(duì)應(yīng)的坐標(biāo)點(diǎn)
  CvPoint* point = (CvPoint*)cvGetSeqElem( seq, idx );
  //定義直線的兩個(gè)端點(diǎn)
  CvPoint line_end[2] = {{0,0}, {0,0}};
  float a, b;
  //累加器的地址指針，也就是霍夫空間的地址指針
  int* adata = (int*)accum.data;
  int i, j, k, x0, y0, dx0, dy0, xflag;
  int good_line;
  const int shift = 16;
  //提取出坐標(biāo)點(diǎn)的橫、縱坐標(biāo)
  i = point->y;
  j = point->x;
 
  // "remove" it by overriding it with the last element
  //用序列中的最后一個(gè)元素覆蓋掉剛才提取出來的隨機(jī)坐標(biāo)點(diǎn)
  *point = *(CvPoint*)cvGetSeqElem( seq, count-1 );
 
  // check if it has been excluded already (i.e. belongs to some other line)
  //檢測(cè)這個(gè)坐標(biāo)點(diǎn)是否已經(jīng)計(jì)算過，也就是它已經(jīng)屬于其他直線
  //因?yàn)橛?jì)算過的坐標(biāo)點(diǎn)會(huì)在掩碼矩陣mask的相對(duì)應(yīng)位置清零
  if( !mdata0[i*width + j] ) //該坐標(biāo)點(diǎn)被處理過
   continue; //不做任何處理，繼續(xù)主循環(huán)
 
  // update accumulator, find the most probable line
  //步驟2，更新累加器矩陣，找到最有可能的直線
  for( n = 0; n < numangle; n++, adata += numrho )
  {
   //由角度計(jì)算距離
   r = cvRound( j * ttab[n*2] + i * ttab[n*2+1] );
   r += (numrho - 1) / 2;
   //在累加器矩陣的相應(yīng)位置上數(shù)值加1，并賦值給val
   int val = ++adata[r];
   //更新最大值，并得到它的角度
   if( max_val < val )
   {
    max_val = val;
    max_n = n;
   }
  }
 
  // if it is too "weak" candidate, continue with another point
  //步驟3，如果上面得到的最大值小于閾值，則放棄該點(diǎn)，繼續(xù)下一個(gè)點(diǎn)的計(jì)算
  if( max_val < threshold )
   continue;
 
  // from the current point walk in each direction
  // along the found line and extract the line segment
  //步驟4，從當(dāng)前點(diǎn)出發(fā)，沿著它所在直線的方向前進(jìn)，直到達(dá)到端點(diǎn)為止
  a = -ttab[max_n*2+1]; //a=-sinθ
  b = ttab[max_n*2]; //b=cosθ
  //當(dāng)前點(diǎn)的橫、縱坐標(biāo)值
  x0 = j;
  y0 = i;
  //確定當(dāng)前點(diǎn)所在直線的角度是在45度～135度之間，還是在0～45或135度～180度之間
  if( fabs(a) > fabs(b) ) //在45度～135度之間
  {
   xflag = 1; //置標(biāo)識(shí)位，標(biāo)識(shí)直線的粗略方向
   //確定橫、縱坐標(biāo)的位移量
   dx0 = a > 0 ? 1 : -1; 
   dy0 = cvRound( b*(1 << shift)/fabs(a) );
   //確定縱坐標(biāo)
   y0 = (y0 << shift) + (1 << (shift-1));
  }
  else //在0～45或135度～180度之間
  {
   xflag = 0; //清標(biāo)識(shí)位
   //確定橫、縱坐標(biāo)的位移量
   dy0 = b > 0 ? 1 : -1;
   dx0 = cvRound( a*(1 << shift)/fabs(b) );
   //確定橫坐標(biāo)
   x0 = (x0 << shift) + (1 << (shift-1));
  }
  //搜索直線的兩個(gè)端點(diǎn)
  for( k = 0; k < 2; k++ )
  {
   //gap表示兩條直線的間隙，x和y為搜索位置，dx和dy為位移量
   int gap = 0, x = x0, y = y0, dx = dx0, dy = dy0;
   //搜索第二個(gè)端點(diǎn)的時(shí)候，反方向位移
   if( k > 0 )
    dx = -dx, dy = -dy;
 
   // walk along the line using fixed-point arithmetics,
   // stop at the image border or in case of too big gap
   //沿著直線的方向位移，直到到達(dá)圖像的邊界或大的間隙為止
   for( ;; x += dx, y += dy )
   {
    uchar* mdata;
    int i1, j1;
    //確定新的位移后的坐標(biāo)位置
    if( xflag )
    {
     j1 = x;
     i1 = y >> shift;
    }
    else
    {
     j1 = x >> shift;
     i1 = y;
    }
    //如果到達(dá)了圖像的邊界，停止位移，退出循環(huán)
    if( j1 < 0 || j1 >= width || i1 < 0 || i1 >= height )
     break;
    //定位位移后掩碼矩陣位置
    mdata = mdata0 + i1*width + j1;
 
    // for each non-zero point:
    // update line end,
    // clear the mask element
    // reset the gap
    //該掩碼不為0，說明該點(diǎn)可能是在直線上
    if( *mdata ) 
    {
     gap = 0; //設(shè)置間隙為0
     //更新直線的端點(diǎn)位置
     line_end[k].y = i1;
     line_end[k].x = j1;
    }
    //掩碼為0，說明不是直線，但仍繼續(xù)位移，直到間隙大于所設(shè)置的閾值為止
    else if( ++gap > lineGap ) //間隙加1
     break;
   }
  }
  //步驟5，由檢測(cè)到的直線的兩個(gè)端點(diǎn)粗略計(jì)算直線的長度
  //當(dāng)直線長度大于所設(shè)置的閾值時(shí)，good_line為1，否則為0
  good_line = abs(line_end[1].x - line_end[0].x) >= lineLength ||
     abs(line_end[1].y - line_end[0].y) >= lineLength;
  //再次搜索端點(diǎn)，目的是更新累加器矩陣和更新掩碼矩陣，以備下一次循環(huán)使用
  for( k = 0; k < 2; k++ )
  {
   int x = x0, y = y0, dx = dx0, dy = dy0;
 
   if( k > 0 )
    dx = -dx, dy = -dy;
 
   // walk along the line using fixed-point arithmetics,
   // stop at the image border or in case of too big gap
   for( ;; x += dx, y += dy )
   {
    uchar* mdata;
    int i1, j1;
 
    if( xflag )
    {
     j1 = x;
     i1 = y >> shift;
    }
    else
    {
     j1 = x >> shift;
     i1 = y;
    }
 
    mdata = mdata0 + i1*width + j1;
 
    // for each non-zero point:
    // update line end,
    // clear the mask element
    // reset the gap
    if( *mdata )
    {
     //if語句的作用是清除那些已經(jīng)判定是好的直線上的點(diǎn)對(duì)應(yīng)的累加器的值，避免再次利用這些累加值
     if( good_line ) //在第一次搜索中已經(jīng)確定是好的直線
     {
      //得到累加器矩陣地址指針
      adata = (int*)accum.data;
      for( n = 0; n < numangle; n++, adata += numrho )
      {
       r = cvRound( j1 * ttab[n*2] + i1 * ttab[n*2+1] );
       r += (numrho - 1) / 2;
       adata[r]--; //相應(yīng)的累加器減1
      }
     }
     //搜索過的位置，不管是好的直線，還是壞的直線，掩碼相應(yīng)位置都清0，這樣下次就不會(huì)再重復(fù)搜索這些位置了，從而達(dá)到減小計(jì)算邊緣點(diǎn)的目的
     *mdata = 0;
    }
    //如果已經(jīng)到達(dá)了直線的端點(diǎn)，則退出循環(huán)
    if( i1 == line_end[k].y && j1 == line_end[k].x )
     break;
   }
  }
  //如果是好的直線
  if( good_line )
  {
   CvRect lr = { line_end[0].x, line_end[0].y, line_end[1].x, line_end[1].y };
   //把兩個(gè)端點(diǎn)壓入序列中
   cvSeqPush( lines, &lr );
   //如果檢測(cè)到的直線數(shù)量大于閾值，則退出該函數(shù)
   if( lines->total >= linesMax )
    return;
  }
 }
}

#include "opencv2/core/core.hpp"
#include "opencv2/highgui/highgui.hpp"
#include "opencv2/imgproc/imgproc.hpp"
 
#include 
using namespace cv;
using namespace std;
 
int main( int argc, char** argv )
{
 Mat src, edge,color_edge;
 src=imread("building.jpg");
 if( !src.data ) 
 return -1; 
 
 Canny(src,edge,50,200,3);
 cvtColor( edge, color_edge, CV_GRAY2BGR );
 vector lines;
 HoughLinesP(edge, lines, 1, CV_PI/180, 80, 30, 10 );
 for( size_t i = 0; i < lines.size(); i++ )
 {
  Vec4i l = lines[i];
  line( color_edge, Point(l[0], l[1]), Point(l[2], l[3]), Scalar(0,0,255), 2);
 }
 
 namedWindow( "lines", CV_WINDOW_AUTOSIZE );
 imshow( "lines", color_edge );
 waitKey(0);
 
 return 0;
}