【原】训练自己的haar-like特征分类器并识别物体(3)
在前两篇文章中,我介绍了《训练自己的haar-like特征分类器并识别物体》的前三个步骤:
1.准备训练样本图片,包括正例及反例样本
2.生成样本描述文件
3.训练样本
4.目标识别
==============
本文将着重说明最后一个阶段——目标识别,也即利用前面训练出来的分类器文件(.xml文件)对图片中的物体进行识别,并在图中框出在该物体。由于逻辑比较简单,这里直接上代码:
int _tmain(int argc, _TCHAR* argv[]) { char *cascade_name = CASCADE_HEAD_MY; //上文最终生成的xml文件命名为"CASCADE_HEAD_MY.xml" cascade = (CvHaarClassifierCascade*)cvLoad( cascade_name, 0, 0, 0 ); //加载xml文件 if( !cascade ) { fprintf( stderr, "ERROR: Could not load classifier cascade\n" ); system("pause"); return -1; } storage = cvCreateMemStorage(0); cvNamedWindow( "face", 1 ); const char* filename = "(12).bmp"; IplImage* image = cvLoadImage( filename, 1 ); if( image ) { detect_and_draw( image ); //函数见下方 cvWaitKey(0); cvReleaseImage( &image ); } cvDestroyWindow("result"); return 0; }
1 void detect_and_draw(IplImage* img ) 2 { 3 double scale=1.2; 4 static CvScalar colors[] = { 5 {{0,0,255}},{{0,128,255}},{{0,255,255}},{{0,255,0}}, 6 {{255,128,0}},{{255,255,0}},{{255,0,0}},{{255,0,255}} 7 };//Just some pretty colors to draw with 8 9 //Image Preparation 10 // 11 IplImage* gray = cvCreateImage(cvSize(img->width,img->height),8,1); 12 IplImage* small_img=cvCreateImage(cvSize(cvRound(img->width/scale),cvRound(img->height/scale)),8,1); 13 cvCvtColor(img,gray, CV_BGR2GRAY); 14 cvResize(gray, small_img, CV_INTER_LINEAR); 15 16 cvEqualizeHist(small_img,small_img); //直方图均衡 17 18 //Detect objects if any 19 // 20 cvClearMemStorage(storage); 21 double t = (double)cvGetTickCount(); 22 CvSeq* objects = cvHaarDetectObjects(small_img, 23 cascade, 24 storage, 25 1.1, 26 2, 27 0/*CV_HAAR_DO_CANNY_PRUNING*/, 28 cvSize(30,30)); 29 30 t = (double)cvGetTickCount() - t; 31 printf( "detection time = %gms\n", t/((double)cvGetTickFrequency()*1000.) ); 32 33 //Loop through found objects and draw boxes around them 34 for(int i=0;i<(objects? objects->total:0);++i) 35 { 36 CvRect* r=(CvRect*)cvGetSeqElem(objects,i); 37 cvRectangle(img, cvPoint(r->x*scale,r->y*scale), cvPoint((r->x+r->width)*scale,(r->y+r->height)*scale), colors[i%8]); 38 } 39 for( int i = 0; i < (objects? objects->total : 0); i++ ) 40 { 41 CvRect* r = (CvRect*)cvGetSeqElem( objects, i ); 42 CvPoint center; 43 int radius; 44 center.x = cvRound((r->x + r->width*0.5)*scale); 45 center.y = cvRound((r->y + r->height*0.5)*scale); 46 radius = cvRound((r->width + r->height)*0.25*scale); 47 cvCircle( img, center, radius, colors[i%8], 3, 8, 0 ); 48 } 49 50 cvShowImage( "result", img ); 51 cvReleaseImage(&gray); 52 cvReleaseImage(&small_img); 53 }
===================================
其实上面的代码可以运用于大部分模式识别问题,无论是自己生成的xml文件还是opencv自带的xml文件。在opencv的工程目录opencv\data文件夹下有大量的xml文件,这些都是opencv开源项目中的程序员们自己训练出来的。然而,效果一般不会合你预期,所以才有了本系列文章。天下没有免费的午餐,想要获得更高的查准率与查全率,不付出点努力是不行的!