手势识别 opencv 代码

      从csdn下载了一个pdf文档，是关于手势识别的。   我把它弄了 下来，并调试好了。  环境：vs2010   opencv 2.4.7

main.cpp

#include <highgui.h>#include <cv.h>#include <vector>using namespace cv;using namespace std;int pmsf_value = 5;//均值飘逸分割平滑系数int MopEx_value = 2;//开运算int Hmatch_value = 25;//模板匹配系数//亮度int V_low = 30;int V_high = 250;//饱和度int S_low = 10;int S_high = 170;//色相int H_low_max = 40;int H_high_min = 100;int if_high_light = 1; //是否高光补偿IplImage *src = 0;IplImage *srcResize = 0;IplImage *img_YCrCb = 0;CvSize newSize;CvSize sz;IplImage *tmp1;IplImage *tmp2;IplImage *tmp3;IplImage *src2;IplImage *src1;////////////////////////////////////////////////////////////////////////////////////////////////////////////////IplImage *YCrCb;IplImage *YCrCb_mask;IplImage *Y_channel,*Cr_channel, *Cb_channel;IplImage *Y_cmp,*Cr_cmp, *Cb_cmp;CvScalar Y_lower;CvScalar Y_upper;CvScalar Cr_lower;CvScalar Cr_upper;CvScalar Cb_lower;CvScalar Cb_upper;CvScalar YCrCb_lower;CvScalar YCrCb_upper;///////////////////////////////////////////////////////////////////////////////////////////////////////////////void init_hand_YCrCb(){//img_YCrCb = cvCreateImage( sz, 8, 3);YCrCb_mask = cvCreateImage(sz,IPL_DEPTH_8U,1);;//最终的图片YCrCb = cvCreateImage(sz,IPL_DEPTH_8U,3);//三通道Y_channel = cvCreateImage(sz,IPL_DEPTH_8U,1);Cr_channel = cvCreateImage(sz,IPL_DEPTH_8U,1);Cb_channel = cvCreateImage(sz,IPL_DEPTH_8U,1);//按范围截取后Y_cmp = cvCreateImage(sz,IPL_DEPTH_8U,1);Cr_cmp = cvCreateImage(sz,IPL_DEPTH_8U,1);Cb_cmp = cvCreateImage(sz,IPL_DEPTH_8U,1);//Y,Cr,Cb 的颜色范围Y_lower = CV_RGB(0,0,130);Y_upper = CV_RGB(0,0,130);Cr_lower = CV_RGB(0,0,125);Cr_upper = CV_RGB(0,0,125);Cb_lower = CV_RGB(0,0,132);Cb_upper = CV_RGB(0,0,147);YCrCb_lower = cvScalar(0,0,132,0);YCrCb_upper = cvScalar(130,125,147,0);}////////////////////////////////////////////////////////////////////////////////////////////////////////////////void hand_YCrCb(){//转换到YCrBrcvCvtColor(src2,img_YCrCb, CV_RGB2YCrCb);//分割到Y,Cr,CbcvSplit(img_YCrCb,Y_channel,Cr_channel,Cb_channel,0);//将Y_channel 的位于 Y_lower 和 Y_upper 之间的元素复制到 Y_tmp 中cvInRangeS( Y_channel, Y_lower, Y_upper, Y_cmp);cvInRangeS(Cr_channel,Cr_lower,Cr_upper ,Cr_cmp);cvInRangeS(Cb_channel,Cb_lower,Cb_upper ,Cb_cmp);//合并Y,Cr,Cb 通道到YCrCb 中cvMerge(Y_cmp,Cr_cmp,Cb_cmp,0,YCrCb);//显示结果cvShowImage("YCrCb_mask",YCrCb);//cvInRangeS (img_YCrCb, YCrCb_lower, YCrCb_upper, YCrCb_mask);//cvShowImage( "YCrCb_mask", YCrCb_mask);}//////////////////////////////////////////////////////////////////////////////////////////////////////////////IplImage* hsv_image;IplImage* hsv_mask;CvScalar hsv_min;CvScalar hsv_max;IplImage *H_img,*S_img, *V_img;IplImage *H_mask, *H_mask1,*S_mask, *S_mask1, *V_mask, *V_mask1,*V_mask2;///////////////////////////////////////////////////////////////////////////////////////////////////////////////void init_hand_HSV(){hsv_image = cvCreateImage( sz, 8, 3);hsv_mask = cvCreateImage( sz, 8, 1);hsv_min = cvScalar(0, 20, 20, 0);hsv_max = cvScalar(20, 250, 255, 0);//hsv_mask->origin = 1;//方法2: 单独处理各个通道H_img = cvCreateImage(sz,IPL_DEPTH_8U,1);S_img = cvCreateImage(sz,IPL_DEPTH_8U,1);V_img = cvCreateImage(sz,IPL_DEPTH_8U,1);H_mask = cvCreateImage(sz,IPL_DEPTH_8U,1);H_mask1 = cvCreateImage(sz,IPL_DEPTH_8U,1);S_mask = cvCreateImage(sz,IPL_DEPTH_8U,1);S_mask1 = cvCreateImage(sz,IPL_DEPTH_8U,1);V_mask = cvCreateImage(sz,IPL_DEPTH_8U,1);V_mask2 = cvCreateImage(sz,IPL_DEPTH_8U,1);V_mask1 = cvCreateImage(sz,IPL_DEPTH_8U,1);}//////////////////////////////////////////////////////////////////////////////////////////////////////////////void color_blance(){CvScalar avg = cvAvg(H_img, 0);printf("%f, %f, %f, %f\n",avg.val[0],avg.val[1],avg.val[2],avg.val[3]);double d = 128 - avg.val[0];avg.val[0] = d;cvAddS(H_img, avg, H_img, 0);}///////////////////////////////////////////////////////////////////////////////////////////////////////////////void hand_HSV(){cvCvtColor(src2, hsv_image, CV_BGR2HSV);//cvInRangeS (hsv_image, hsv_min, hsv_max, hsv_mask);//cvShowImage( "hsv_msk", hsv_mask);//方法2: 单独处理各个通道cvSplit(hsv_image,H_img,S_img,V_img,0);//color_blance();//cvMerge(H_img,S_img,V_img,0,hsv_image);//cvShowImage( "色彩平衡后", hsv_image);//cvShowImage( "H 通道", H_img);//直方图均衡化(效果更差)//cvEqualizeHist(H_img, H_img);//cvShowImage( "H 通道_均衡化", H_img);//自适应//cvAdaptiveThreshold(H_img, H_mask, 30, 0, 0, 3, 5);//cvShowImage( "H", H_img);//cvShowImage( "S", S_img);//cvShowImage( "V", V_img);//色相cvInRangeS(H_img,cvScalar(0,0,0,0),cvScalar(H_low_max,0,0,0),H_mask);//红色区cvInRangeS(H_img,cvScalar(256 -H_high_min,0,0,0),cvScalar(256,0,0,0),H_mask1);//紫色区//饱和度cvInRangeS(S_img,cvScalar(S_low,0,0,0),cvScalar(S_high,0,0,0),S_mask); //中间区//cvInRangeS(S_img,cvScalar(20,0,0,0),cvScalar(100,0,0,S_mask1); //低饱和度//亮度cvInRangeS(V_img,cvScalar(V_high,0,0,0),cvScalar(256,0,0,0),V_mask);//高亮区cvInRangeS(V_img,cvScalar(V_low,0,0,0),cvScalar(V_high,0,0,0),V_mask1); //中间区//cvInRangeS(V_img,cvScalar(150,0,0,0),cvScalar(250,0,0,0),V_mask2); //较亮区//红黄, 和蓝紫的混合cvOr(H_mask1, H_mask, H_mask, 0);//消除饱和度过低区域cvAnd(H_mask,S_mask,H_mask,0);//cvShowImage( "饱和度过滤", H_mask);//消去过亮过暗区域cvAnd(H_mask,V_mask1,H_mask,0);//cvShowImage( "亮度过滤", H_mask);//cvShowImage( "hsv_msk", H_mask);//补偿过亮区域if(if_high_light) cvOr(H_mask,V_mask,H_mask,0);//cvShowImage( "补偿高光", H_mask);//cvShowImage( "曝光过度 V", V_mask);//cvShowImage( "曝光补偿", S_mask);//是否补偿曝光过度hsv_mask = H_mask;//cvShowImage( "hsv_msk", H_mask);}/////////////////////////////////////////////////////////////////////////////////////////////////////////////////阀值化IplImage* thd_src;IplImage* thd_dst1;IplImage* thd_dst2;int thd_max = 255;int thd_val = 100;void inti_threshold(){thd_src = cvCreateImage(sz,IPL_DEPTH_8U,1);thd_dst1 = cvCreateImage(sz,IPL_DEPTH_8U,1);thd_dst2 = cvCreateImage(sz,IPL_DEPTH_8U,1);}void threshold(){cvCvtColor(src1, thd_src, CV_RGB2GRAY);cvAdaptiveThreshold(thd_src, thd_dst1, thd_max,CV_ADAPTIVE_THRESH_MEAN_C, CV_THRESH_BINARY, 3, 5);cvThreshold(thd_src, thd_dst2, thd_val, thd_max,CV_THRESH_BINARY);cvShowImage( "阀值前", thd_src);cvShowImage( "阀值化1", thd_dst1);cvCreateTrackbar( "thd_max", "阀值化1", &thd_max, 256, 0);cvShowImage( "阀值化2", thd_dst2);cvCreateTrackbar( "thd_val", "阀值化2", &thd_val, 256, 0);}///////////////////////////////////////////////////////////////////////////////////////////////////////////////void resizeSrc(){double scale = 0.5;//获得图像大小sz = cvGetSize(src);newSize.height = (int)(sz.height * scale);newSize.width = (int)(sz.width * scale);src = cvCreateImage(newSize,IPL_DEPTH_8U,3);cvResize(src, src, CV_INTER_LINEAR);}///////////////////////////////////////////////////////////////////////////////////////////////////////////////IplImage* smooth1;IplImage* smooth2;IplImage* smooth3;IplImage* smooth4;IplImage* smooth5;void reduce_noise(){//cvSmooth(hsv_mask, smooth1, CV_GAUSSIAN, 3, 0, 0, 0);//cvSmooth(hsv_mask, smooth1, CV_MEDIAN, 3, 0, 0, 0);//cvSmooth(smooth1, smooth1, CV_BILATERAL, 3, 0, 0, 0);//cvDilate(hsv_mask, smooth1, 0 ,2);//cvErode(smooth1 ,smooth2, 0, 2);cvMorphologyEx(hsv_mask, smooth1, 0, NULL, CV_MOP_CLOSE,MopEx_value);//调试之前的cvMorphologyEx(hsv_mask, smooth1, 0, CV_SHAPE_RECT, CV_MOP_CLOSE,MopEx_value);    //cvMorphologyEx(smooth1, smooth2, 0, CV_SHAPE_RECT, CV_MOP_OPEN,1);cvShowImage( "扩张腐蚀", smooth1);//cvSmooth(smooth2, smooth3, CV_MEDIAN, 9, 0, 0, 0);//cvShowImage( "平滑", smooth3);}//////////////////////////////////////////////////////////////////////////////////////////////////////////////////轮廓匹配IplImage* g_image = NULL;IplImage* g_gray = NULL;int g_thresh = 100;CvMemStorage* g_storage = NULL;CvMemStorage* g_storage1 = NULL;CvMemStorage* g_storage2 = NULL;CvSeq* seqMidObj = 0;//塞选后的轮廓集合int handNum = 0;///////////////////////////////////////////////////////////////////////////void hand_contours(IplImage* dst) {if( g_storage==NULL ) {g_storage = cvCreateMemStorage(0);g_storage1 = cvCreateMemStorage(0);g_storage2 = cvCreateMemStorage(0);} else {cvClearMemStorage( g_storage );cvClearMemStorage( g_storage1 );cvClearMemStorage( g_storage2 );}int i = 0, j=0;CvSeq* contours = 0;CvSeq* contoursHead = 0;CvSeq* p = NULL;CvSeq* q = NULL;seqMidObj = 0;handNum = 0;//cvThreshold( g_gray, g_gray, g_thresh, 255, CV_THRESH_BINARY );cvFindContours( dst, g_storage, &contours, sizeof(CvContour),CV_RETR_EXTERNAL);//只查找外轮廓contoursHead = contours;//contours 的头//cvZero( g_gray );/*CvSeq* seqApprox;CvSeq* head;int j=0;*//*for(p=contours;p != NULL; p = p->h_next){j++;seqApprox = cvApproxPoly(p, sizeof(CvContour),g_storage1,CV_POLY_APPROX_DP, 2, 0);seqApprox = seqApprox->h_next;if(p==contours) head = seqApprox;}printf("total = %d\n ",j );seqApprox = head;*/cvZero( tmp3 );//cvZero( tmp2 );if( contours )cvDrawContours( tmp3, contours,cvScalarAll(255),cvScalar(255,0,0,0),1);//绘制轮廓//if( contours )cvDrawContours( tmp2, seqApprox,cvScalar(255,0,0,0),cvScalar(255,100,0,0),1);//绘制轮廓//cvShowImage( "查找轮廓", tmp3);//CvSeq* seqMidObj = 0;// = cvCreateSeq(CV_SEQ_ELTYPE_CODE,sizeof(CvSeq), sizeof(int),g_storage2);//去除与窗口邻接的轮廓contours = contoursHead;i=0;CvRect bound;int dat = 2;//去除小面积区域int contArea = 0;int imgArea = newSize.height * newSize.width;for( ; contours != 0; contours = contours->h_next ){i++;//如果面积过小, 则排除contArea = fabs(cvContourArea( contours, CV_WHOLE_SEQ ));if ( (double)contArea/imgArea < 0.015 ){continue;}//如果边界与窗口相连, 则排除bound = cvBoundingRect(contours, 0);if( bound.x < dat||bound.y < dat||bound.x + bound.width + dat > newSize.width||bound.y + bound.height + dat > newSize.height){//printf(" %d, %d, %d, %d\n",bound.x, bound.y,bound.width, bound.height );//cvRectangle(tmp3, cvPoint(bound.x,bound.y),cvPoint(bound.x + bound.width, bound.y +bound.height),cvScalar(255,255,255,255),1,8,0);continue;}//建立轮廓链表q = p;//p = cvCloneSeq(contours, g_storage2);p = contours;if(q == NULL){seqMidObj = p;//p->h_next = NULL;//p->h_prev = NULL;//printf("第1 个!");}else{q->h_next = p;p->h_prev = q;//printf("1 个!");}//j++;handNum++;}//printf("找到轮廓: %d 个 塞选: %d 个\n", i,j);if(seqMidObj){seqMidObj->h_prev = NULL;p->h_next = NULL;}if(handNum>0) printf("找到手: %d ",handNum);//CvSeq* seqMidObj_head = seqMidObj;cvZero( tmp3 );if( seqMidObj )cvDrawContours( tmp3, seqMidObj,cvScalarAll(255),cvScalar(255,0,0,0),1);//绘制轮廓//cvShowImage( "轮廓塞选", tmp3);}/////////////////////////////////////////////////////////////////////////////////////////////////////////////IplImage* tmp_img = 0;CvMemStorage* storage_tmp = 0;CvSeq* handT = 0;CvSeq* handT1 = 0;CvSeq* handT2 = 0;int handTNum = 10;//10 个模板char *tmp_names[] ={"1.bmp","2.bmp","3.bmp","4.bmp","5.bmp","6.bmp","7.bmp","8.bmp","9.bmp","10.bmp"};char *num_c[] = {"1", "2", "3", "4", "5", "6", "7", "8", "9", "10"};///////////////////////////////////////////////////////载入模板的轮廓void init_hand_template(){storage_tmp = cvCreateMemStorage(0);int i = 0;for(i=0; i<handTNum; i++){tmp_img = cvLoadImage(tmp_names[i],CV_LOAD_IMAGE_GRAYSCALE);if(!tmp_img){printf("未找到文件: %s\n", tmp_names[i]);continue;}//cvShowImage("载入模板", tmp_img);handT1 = handT2;cvFindContours( tmp_img, storage_tmp, &handT2,sizeof(CvContour), CV_RETR_EXTERNAL);if(handT2){printf("载入模板: %s 成功!\n",tmp_names[i]);if(handT1 == NULL){printf("载入第一个模板!\n");handT = handT2;}else{handT2->h_prev = handT1;handT1->h_next = handT2;}}}}/////////////////////////////////////////////////////////////////////////////////////////////////////////////bool if_match_num = false;int match_num = -1;//模板匹配手void hand_template_match(CvSeq* handT, CvSeq* hand){int i=0;int kind = -1;double hu = 1;double hutmp;CvSeq* handp = handT;int method = CV_CONTOURS_MATCH_I1;match_num = 0;if(handT==NULL){return;printf("handT==NULL!\n");}if(hand==NULL){return;printf("hand==NULL!\n");}for(i=0; i<handTNum; i++){hutmp = cvMatchShapes(handp, hand, method, 0);handp = handp->h_next;//找到hu 矩最小的模板if(hu > hutmp){hu = hutmp;kind = i+1;}//printf("%f ", hu);}//显示匹配结果if(hu<((double)Hmatch_value)/100){printf("匹配模板: %d (%f)", kind, hu);match_num = kind;if_match_num = true;}else{if_match_num = false;}}////////////////////////////////////////////////////////////////////////////////////////////////////////////////拉普拉斯变换int sigma = 2;int smoothType = CV_GAUSSIAN;IplImage* laplace = 0;IplImage* colorlaplace = 0;IplImage* planes[3] = { 0, 0, 0 };int ls_low = 247;void init_laplace(){int i;for( i = 0; i < 3; i++ )planes[i] = cvCreateImage(newSize, 8, 1 );laplace = cvCreateImage( newSize, IPL_DEPTH_16S, 1 );colorlaplace = cvCreateImage( newSize, 8, 3 );}void toLaplace(IplImage* dst){int i, c, ksize;ksize = (sigma*5)|1;//cvSmooth( dst, colorlaplace, smoothType, ksize, ksize, sigma,sigma );cvSplit( dst, planes[0], planes[1], planes[2], 0 );for( i = 0; i < 3; i++ ){cvLaplace( planes[i], laplace, 5 );cvConvertScaleAbs( laplace, planes[i], (sigma+1)*0.25, 0 );}cvMerge( planes[0], planes[1], planes[2], 0, colorlaplace );colorlaplace->origin = dst->origin;//前期平滑//cvMorphologyEx(planes[2], planes[2], 0, CV_SHAPE_RECT,CV_MOP_CLOSE, 1);cvShowImage("拉普拉斯变换", colorlaplace );//smoothType = smoothType == CV_GAUSSIAN ? CV_BLUR : smoothType== CV_BLUR ? CV_MEDIAN : CV_GAUSSIAN;cvCvtColor(colorlaplace, colorlaplace, CV_BGR2HSV);cvSplit(colorlaplace, planes[0], planes[1], planes[2], 0);cvInRangeS (planes[2], cvScalar(ls_low,0,0,0),cvScalar(256,0,0,0), planes[2]);//后期平滑//cvMorphologyEx(planes[2], planes[2], 0, CV_SHAPE_RECT,CV_MOP_CLOSE, 1);//cvSmooth(planes[2], planes[2], CV_MEDIAN, 3, 0, 0, 0);//cvErode(planes[2] ,planes[2], 0, 1);cvShowImage("拉普拉斯_边缘化", planes[2]);}///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////#define UP -1#define DOWN 1#define LEFT 1#define RIGHT -1CvPoint hand_center = cvPoint(0,0); //按外接矩形求中心点CvPoint hand_center_last = cvPoint(0,0); //CvPoint hand_direct_now = cvPoint(0,0); //本次监测到的移动方向CvPoint hand_direct_last = cvPoint(0,0); //上次监测到的移动方向CvPoint hand_direct = cvPoint(0,0); //合成的方向//求手的中点和移动方向void hand_direction(CvSeq* hand){//如果没有手, 则清除结果if(!hand){hand_center = cvPoint(0,0);hand_center_last = cvPoint(0,0);hand_direct_now = cvPoint(0,0);hand_direct_last = cvPoint(0,0);hand_direct = cvPoint(0,0);return;}hand_center_last = hand_center;//获得中心点CvRect bound = cvBoundingRect(hand, 0);hand_center.x = bound.x + bound.width/2;hand_center.y = bound.y + bound.height/2;if(hand_center_last.x!=0){hand_direct_now.x = hand_center.x - hand_center_last.x;hand_direct_now.y = hand_center.y - hand_center_last.y;if(hand_direct_now.x!=0){hand_direct.x = (hand_direct_now.x +hand_direct_last.x)/2;if(hand_direct.x!=0){printf(" X 移动: %d ",hand_direct.x);}}if(hand_direct_now.y!=0){hand_direct.y = (hand_direct_now.y +hand_direct_last.y)/2;if(hand_direct.y!=0){printf(" Y 移动: %d ",hand_direct.y);}}}}///////////////////////////////////////////////////////////////////////////////////////////////////////////////绘制识别结果void hand_draw(IplImage* dst,CvSeq* hands ){if(!hands) return;CvRect bound;//凸包int i, hullcount;CvPoint pt0;CvSeq* hull;CvSeq* handp = hands;//凸包缺陷CvConvexityDefect* defect;CvSeq* hullDefect;//CvSeq* hullDefectSelect;int hullDefectNum=0;//cvPoint** points = (cvPoint*)malloc(sizeof(cvPoint)*3);//cvPoint points[3];//绘制轮廓cvDrawContours( dst, handp,cvScalar(255,150,100,0),cvScalar(255,0,0,0),1, 1, 8,cvPoint(0,0));CvFont font;//绘制检测到的数字if(if_match_num){//外阴影cvInitFont(&font,CV_FONT_HERSHEY_SIMPLEX,1.0f,1.0f,0,5,8);cvPutText(dst, num_c[match_num-1], cvPoint(5, 30), &font,CV_RGB(255,255,255));//内颜色cvInitFont(&font,CV_FONT_HERSHEY_SIMPLEX,1.0f,1.0f,0,2,8);cvPutText(dst, num_c[match_num-1], cvPoint(5, 30), &font,CV_RGB(255,0,0));}//绘制移动方向if(1){//cvInitFont(&font,CV_FONT_HERSHEY_SIMPLEX,1.0f,1.0f,0,//cvPutText(dst, "X: ", cvPoint(5, 30), &font,CV_RGB(255,0,0));}//对提取出的轮廓遍历for( ; handp != 0; handp = handp->h_next ){bound = cvBoundingRect(handp, 0);//求并绘制中心点//内颜色int det = 2;cvRectangle(dst,cvPoint(hand_center.x-det, hand_center.y-det),cvPoint(hand_center.x+det, hand_center.y+det),CV_RGB(0,0,0),3,8,0);//外轮廓det = 3;cvRectangle(dst,cvPoint(hand_center.x-det, hand_center.y-det),cvPoint(hand_center.x+det, hand_center.y+det),CV_RGB(255,255,255),1,8,0);//绘出外包络方框cvRectangle(dst,cvPoint(bound.x, bound.y),cvPoint(bound.x + bound.width, bound.y +bound.height),cvScalar(0,0,255,0),2,8,0);//寻找凸包hull = cvConvexHull2( handp, 0, CV_CLOCKWISE, 0 );hullcount = hull->total;//printf("凸包点数: %d ",hullcount);pt0 = **CV_GET_SEQ_ELEM( CvPoint*, hull, hullcount - 1 );//画凸包for( i = 0; i < hullcount; i++ ){//得到凸包的点CvPoint pt = **CV_GET_SEQ_ELEM( CvPoint*, hull, i );cvLine( dst, pt0, pt, CV_RGB( 0, 255, 0 ), 1, CV_AA, 0 );pt0 = pt;}//检查缺陷/*if(!cvCheckContourConvexity(hands)){hullDefect = cvConvexityDefects(hands, hull, 0);hullDefectNum = hullDefect->total;printf("缺陷个数: %d ",hullDefectNum);for( i = 0; i < hullDefectNum; i++ ){defect =(CvConvexityDefect*)cvGetSeqElem(hullDefect, i);cvLine( dst,cvPoint(defect->end->x,defect->end->y), cvPoint(defect->depth_point->x,defect->depth_point->y), CV_RGB(150,150,150),1,8,0);cvLine( dst,cvPoint(defect->start->x,defect->start->y),cvPoint(defect->depth_point->x,defect->depth_point->y), CV_RGB(150,150,150),1,8,0);}}*/}}/////////////////////////////////////////////////////////////////////////////////////////////////////////////void resizeAllWindow(){cvNamedWindow("src",0);cvNamedWindow( "扩张腐蚀",0);cvNamedWindow( "最终识别",0);cvResizeWindow("src",newSize.width,newSize.height);cvNamedWindow( "参数调试",0);cvResizeWindow("参数调试",newSize.width*1.5,60*11);cvCreateTrackbar( "均值漂移滤波", "参数调试", &pmsf_value, 20, 0 );cvCreateTrackbar( "开运算降噪", "参数调试", &MopEx_value, 5, 0 );cvCreateTrackbar( "色相红黄区", "参数调试", &H_low_max, 150, 0 );cvCreateTrackbar( "色相蓝紫区", "参数调试", &H_high_min, 150, 0 );cvCreateTrackbar( "亮度下限", "参数调试", &V_low, 100, 0 );cvCreateTrackbar( "亮度上限", "参数调试", &V_high, 256, 0 );cvCreateTrackbar( "饱和度下限", "参数调试", &S_low, 100, 0 );cvCreateTrackbar( "饱和度上限", "参数调试", &S_high, 255, 0 );cvCreateTrackbar( "高光补偿", "参数调试", &if_high_light, 1, 0 );cvCreateTrackbar( "match 系数", "参数调试", &Hmatch_value, 50, 0 );}/////////////////////////////////////////////////////////////////////////////////////////////////////////////////这段代码转移到 handDlg.cpp//////控制鼠标//void control_mouse()//{// CPoint point;// GetCursorPos(&point);//// int scale = 10;//// point.x += hand_direct.x * scale;// point.y += hand_direct.y * scale;//// SetCursorPos( point.x, point.y);//////}/////////////////////////////////////////////////////////////////////////////////////////////////////////////int main(){int c = 0;double scale = 0.5;// //打开摄像头// CvCapture* capture = cvCaptureFromCAM(0);// src = cvQueryFrame(capture);//// //打开摄像头失败// if(!src){// printf("未找到摄像头");// return NO_CAMERA_BUG;// }//// //获得图像大小// sz = cvGetSize(src);// newSize.height = (int)(sz.height * scale);// newSize.width = (int)(sz.width * scale);const std::string sourceReference = "F:\\HMF\\项目\\服创大赛_物联网\\机器视觉\\素材\\视频\\1.MP4";int delay = 5;//char c;int frameNum = -1; // Frame counterVideoCapture captRefrnc(sourceReference);if ( !captRefrnc.isOpened()){// cout << "Could not open reference " <<sourceReference << endl;return -1;}Size sz = Size( (int) captRefrnc.get(CV_CAP_PROP_FRAME_WIDTH),(int) captRefrnc.get(CV_CAP_PROP_FRAME_HEIGHT) );newSize.height = (int)(sz.height * scale);newSize.width = (int)(sz.width * scale);sz = newSize;//重新调整图像大小//resizeSrc();//建立所有窗体resizeAllWindow();tmp1 = cvCreateImage(sz,IPL_DEPTH_8U,1);tmp2 = cvCreateImage(sz,IPL_DEPTH_8U,3);tmp3 = cvCreateImage(sz,IPL_DEPTH_8U,3);src2 = cvCreateImage(sz,IPL_DEPTH_8U,3);smooth1 = cvCreateImage(sz,IPL_DEPTH_8U,1);smooth2 = cvCreateImage(sz,IPL_DEPTH_8U,1);smooth3 = cvCreateImage(sz,IPL_DEPTH_8U,1);smooth4 = cvCreateImage(sz,IPL_DEPTH_8U,1);smooth5 = cvCreateImage(sz,IPL_DEPTH_8U,1);init_hand_YCrCb();init_hand_HSV();init_laplace();inti_threshold();//载入匹配的模板init_hand_template();src1 = cvCreateImage(sz,IPL_DEPTH_8U,3);/////////////////////开始循环///////////////////////////////while( c != 27){//当前图像// src = cvQueryFrame(capture);captRefrnc >> (Mat)src;//缩小要处理的图像(减小运算量)cvResize(src, src1, CV_INTER_LINEAR);cvShowImage("src",src1);cvPyrMeanShiftFiltering(src1, src2, pmsf_value, 40, 2);//cvSmooth(src1, tmp4, CV_MEDIAN, 9, 0, 0, 0);//cvShowImage( "均值漂移分割", tmp4);//拉普拉斯变换//toLaplace(src2);//hand_YCrCb();//hsv 色彩提取hand_HSV();//阀值化//threshold();//降噪reduce_noise();//寻找手轮廓hand_contours(smooth1);//hand_contours(planes[2]);//hand_contours(tmp2);//处理(未使用)//handProcess(seqMidObj);//寻找匹配hand_template_match(handT, seqMidObj);//求手的移动方向hand_direction(seqMidObj);//绘制额外信息cvZero( tmp2 );hand_draw(tmp2, seqMidObj); //绘制在检测窗口hand_draw(src1, seqMidObj); //绘制在原窗口//cvShowImage( "扩张腐蚀", smooth1);cvShowImage( "最终识别", tmp2 );cvShowImage("src",src1);//控制鼠标(当拳头状, 则控制鼠标移动)//if(match_num == 10) control_mouse(hand_direct.x,hand_direct.y);printf("\n");c = cvWaitKey( 10);}/////////////////////////////////////////////////////////////cvReleaseCapture( &capture);cvDestroyAllWindows();return 0;}

descript.cpp

#include <iostream> // for standard I/O#include <string> // for strings#include <iomanip> // for controlling float print precision#include <sstream> // string to number conversion#include <opencv2/imgproc/imgproc.hpp> // Gaussian Blur#include <opencv2/core/core.hpp> // Basic OpenCV structures(cv::Mat, Scalar)#include <opencv2/highgui/highgui.hpp> // OpenCV window I/Ousing namespace cv;using namespace std;int main(int argc, char *argv[]){const std::string sourceReference = "F:\\HMF\\项目\\服创大赛_物联网\\机器视觉\\素材\\视频\\1.MP4";int delay = 1;char c;int frameNum = -1; // Frame counterVideoCapture captRefrnc(sourceReference);if ( !captRefrnc.isOpened()){// cout << "Could not open reference " <<sourceReference << endl;return -1;}Size refS = Size( ((int) captRefrnc.get(CV_CAP_PROP_FRAME_WIDTH))/4,((int) captRefrnc.get(CV_CAP_PROP_FRAME_HEIGHT))/4);bool bHandFlag = false;const char* WIN_SRC = "Source";const char* WIN_RESULT = "Result";// WindowsnamedWindow(WIN_SRC, CV_WINDOW_AUTOSIZE );namedWindow(WIN_RESULT, CV_WINDOW_AUTOSIZE);Mat frame; // 输入视频帧序列Mat frameHSV; // hsv 空间Mat mask(frame.rows, frame.cols, CV_8UC1); // 2 值掩膜Mat dst(frame); // 输出图像// Mat frameSplit[4];vector< vector<Point> > contours; // 轮廓vector< vector<Point> > filterContours; // 筛选后的轮廓vector< Vec4i > hierarchy; // 轮廓的结构信息vector< Point > hull; // 凸包络的点集while(true) //Show the image captured in the window and repeat{captRefrnc >> frame;resize(frame, frame, refS, 1, 1, INTER_AREA );//改变图像大小if( frame.empty() ){cout << " < < < Game over! > > > ";break;}imshow( WIN_SRC, frame);// Begin// 中值滤波，去除椒盐噪声medianBlur(frame, frame, 5);// GaussianBlur( frame, frameHSV, Size(9, 9), 2, 2 );// imshow("blur2", frameHSV);// pyrMeanShiftFiltering(frame, frameHSV, 10, 10);// imshow(WIN_BLUR, frameHSV);// 转换到HSV 颜色空间，更容易处理cvtColor( frame, frameHSV, CV_BGR2HSV );// split(frameHSV, frameSplit);// imshow(WIN_H, frameSplit[0]);// imshow(WIN_S, frameSplit[1]);// imshow(WIN_V, frameSplit[2]);Mat dstTemp1(frame.rows, frame.cols, CV_8UC1);Mat dstTemp2(frame.rows, frame.cols, CV_8UC1);// 对HSV 空间进行量化，得到2 值图像，亮的部分为手的形状inRange(frameHSV, Scalar(0,30,30), Scalar(40,170,256),dstTemp1);inRange(frameHSV, Scalar(156,30,30), Scalar(180,170,256),dstTemp2);bitwise_or(dstTemp1, dstTemp2, mask);// inRange(frameHSV, Scalar(0,30,30), Scalar(180,170,256),dst);// 形态学操作，去除噪声，并使手的边界更加清晰Mat element = getStructuringElement(MORPH_RECT,Size(3,3));erode(mask, mask, element);morphologyEx(mask, mask, MORPH_OPEN, element);dilate(mask, mask, element);morphologyEx(mask, mask, MORPH_CLOSE, element);frame.copyTo(dst, mask);contours.clear();hierarchy.clear();filterContours.clear();// 得到手的轮廓findContours(mask, contours, hierarchy, CV_RETR_EXTERNAL,CV_CHAIN_APPROX_SIMPLE);// 去除伪轮廓for (size_t i = 0; i < contours.size(); i++){// approxPolyDP(Mat(contours[i]),Mat(approxContours[i]), arcLength(Mat(contours[i]), true)*0.02, true);if (fabs(contourArea(Mat(contours[i]))) >30000)//判断手进入区域的阈值{filterContours.push_back(contours[i]);}}// 画轮廓drawContours(dst, filterContours, -1, Scalar(0,0,255), 3/*,hierarchy*/);// 得到轮廓的凸包络for (size_t j=0; j<filterContours.size(); j++){convexHull(Mat(filterContours[j]), hull, true);int hullcount = (int)hull.size();for (int i=0; i<hullcount-1; i++){line(dst, hull[i+1], hull[i],Scalar(255,0,0), 2, CV_AA);}line(dst, hull[hullcount-1], hull[0],Scalar(255,0,0), 2, CV_AA);}imshow(WIN_RESULT, dst);dst.release();// Endc = cvWaitKey(delay);if (c == 27) break;}}