Opencv实现的利用监控摄像头对高速路口车流量计数.zip

#include<cv.h> #include<highgui.h> #include<cxcore.h> #include<math.h> #include<stdio.h> int main() { IplImage *frame,*pimg,*pimg1; CvCapture *capture; int i,j,t1,t2,t3,n=0,count=0,k1,k2,flag[4]={0},k=0,x1,x2,y1,y2,m,d[4]={70,110,150,185},N=0,sum1=0,sum2=0,sum3=0; int c[8][2]={{0,1},{-1,1},{-1,0},{-1,-1},{0,-1},{1,-1},{1,0},{1,1}};//寻址坐标 int e[8]={1,0,1,0,1,0,1,0}; char *a,**b,*a1,**b1; CvScalar color=cvScalar(255,255,255); capture=cvCreateFileCapture("highwayII_raw.AVI"); pimg=cvLoadImage("backgroud1.bmp",1);//读入背景图像 a=pimg->imageData; b=new char*[pimg->height]; for(i=0;i<pimg->height;i++) { b[i]=&a[i*pimg->widthStep]; //b1[i]=&a1[i*pimg->widthStep]; } cvNamedWindow("test1",CV_WINDOW_AUTOSIZE); cvNamedWindow("test2",CV_WINDOW_AUTOSIZE); while(1)//每次循环完成对一帧图像的处理 {N++; frame=cvQueryFrame(capture); if(!frame)break; cvSaveImage("test2.bmp",frame); pimg1=cvLoadImage("test2.bmp",1); //printf("%d %d %d %d\n",frame->height,frame->widthStep,pimg->height,pimg->widthStep); a1=pimg1->imageData; b1=new char*[pimg1->height]; for(i=0;i<pimg->height;i++) { //b[i]=&a[i*pimg->widthStep]; b1[i]=&a1[i*pimg->widthStep]; } //for(i=0;i<pimg->width;i++) //printf("%d %d %d %d %d %d\n",b[0][3*i],b1[0][3*i],b[0][3*i+1],b1[0][3*i+1],b[0][3*i+2],b1[0][3*i+2]); //领域平均滤波，对图像进行预处理 for(i=1;i<pimg->height-1;i++) { for(j=1;j<pimg->width-1;j++) { sum1=0;sum2=0;sum3=0; for(k=0;k<8;k++)//读入领域像素 { sum1+=unsigned char(b1[i+c[k][0]][3*(j+c[k][1])]*e[k]); sum2+=unsigned char(b1[i+c[k][0]][3*(j+c[k][1])+1]*e[k]); sum3+=unsigned char(b1[i+c[k][0]][3*(j+c[k][1])+2]*e[k]); } b1[i][3*j]=(sum1+unsigned char(b1[i][3*j]))/5; b1[i][3*j+1]=(sum2+unsigned char(b1[i][3*j+1]))/5; b1[i][3*j+2]=(sum2+unsigned char(b1[i][3*j+2]))/5; } } //得到差值图像并用大津法得到的阈值进行运动目标的检测 for(i=0;i<pimg->height;i++) for(j=0;j<pimg->width;j++) { /*t1=(unsigned char)b[i][3*j]-(unsigned char)b1[i][3*j]; t2=(unsigned char)b[i][3*j+1]-(unsigned char)b1[i][3*j+1]; t3=(unsigned char)b[i][3*j+2]-(unsigned char)b1[i][3*j+2];*/ t1=unsigned char(b[i][3*j])-unsigned char(b1[i][3*j]); t2=unsigned char(b[i][3*j+1])-unsigned char(b1[i][3*j+1]); t3=unsigned char(b[i][3*j+2])-unsigned char(b1[i][3*j+2]); //if(i==0) //printf("%d %d %d\n",t1,t2,t3); if(abs(t1)>40&&abs(t2)>40&&abs(t3)>40)//二值化 { b1[i][3*j]=255; b1[i][3*j+1]=255; b1[i][3*j+2]=255; } else { b1[i][3*j]=0; b1[i][3*j+1]=0; b1[i][3*j+2]=0; } } /*for(i=0;i<pimg->height;i++) for(j=0;j<pimg->width;j++) printf("%d,%d,%d\n",b1[i][3*j],b1[i][3*j+1],b1[i][3*j+2]); /*for(i=0;i<pimg->height;i+=40) for(j=0;j<pimg->width;j+=30) { for(k1=0;k1<40;k1++) for(k2=0;k2<30;k2++) if(b1[i+k1][j+3*k2]&&b1[i+k1][j+3*k2+1]&&b1[i+k1][j+3*k2+2]) n++; if(n>210) { x1=i;y1=j; x2=i+40;y2=j+30; CvPoint pt1=cvPoint(x1,y1); CvPoint pt2=cvPoint(x2,y2); cvRectangle(frame,pt1,pt2,color); } n=0; }*/ //车辆计数 m=0;k=0; while(m<4)//对四个检测窗口分别进行检测 {i=d[m]; for(k1=0;k1<10;k1++)//选取窗口的大小为10*30 for(j=0;j<30;j++) { if(b1[50+k1][3*(i+j)]&&b1[50+k1][3*(i+j)+1]&&b1[50+k1][3*(i+j)+2])//像素值不为0说明是目标点 {n++;//printf("%d,%d,%d\n",b1[50][3*(i+j)],b1[50][3*(i+j)+1],b1[50][3*(i+j)+2]); } } //printf("%d\n",n); if(n>35)//如果大于阈值 { if(flag[k]==0)//如果之前没有车进入检测窗口 flag[k]=1;//有车进入检测窗口的标志 }else { if(n<35&&flag[k]==1) { count++;flag[k]=0;printf("%d\n",count);//有车要离开，记录车辆数 } } n=0;m++;k++; } //为了更好的观察实验结果，将四个检测窗口画到图像上对应的位置 CvPoint pt1=cvPoint(70,50); CvPoint pt2=cvPoint(100,60); CvScalar color=cvScalar(255,255,255); cvRectangle(pimg1,pt1,pt2,color); pt1=cvPoint(110,50); pt2=cvPoint(140,60); cvRectangle(pimg1,pt1,pt2,color); pt1=cvPoint(150,50); pt2=cvPoint(180,60); cvRectangle(pimg1,pt1,pt2,color); pt1=cvPoint(185,50); pt2=cvPoint(230,60); cvRectangle(pimg1,pt1,pt2,color); cvShowImage("test2",pimg1); cvShowImage("test1",frame); cvWaitKey(33); cvSaveImage("test1.bmp",pimg1); cvReleaseImage(&pimg1); } cvDestroyWindow("test1"); cvDestroyWindow("test2"); cvReleaseImage(&pimg); return 0; }