基于OpenCV读取摄像头进行人脸检测和人脸识别

前段时间使用OpenCV的库函数实现了人脸检测和人脸识别，笔者的实验环境为VS2010+OpenCV2.4.4，opencv的环境配置网上有很多，不再赘述。检测的代码网上很多，记不清楚从哪儿copy的了，识别的代码是从OpenCV官网上找到的：http://docs.opencv.org/trunk/modules/contrib/doc/facerec/facerec_api.html

需要注意的是，opencv的FaceRecogizer目前有三个类实现了它，特征脸和fisherface方法最少训练图像为两张，而LBP可以单张图像训练。本人的实验采用的图片是100x100大小的，所以如果要添加自己的图像进行识别的话务必调整为100x100，不然会报错。当然在recog_and_draw这个函数里，笔者也将每次检测到的人脸进行了保存，拖出来重命名就可以，路径自己找吧。使用不同的方法识别时，其阈值设置也不同，LBP大概在100，其他两种方法大概在1000。本人的代码已共享，下载链接：http://download.csdn.net/detail/u010944555/6749725

ps：有人说代码的检测率不高，其实可以归结为两方面的原因，第一人脸检测率不高，这个可以通过嵌套检测嘴角、眼睛等来降低，或者背景、光照固定的话可以通过图像差分来解决；第二是识别方法本身的问题，如果想提高识别率，可以添加多张不同姿态、光照下的人脸作为训练的样本，如果有时间的话可以在采集图像时给出一个人脸框，引导用户对齐人脸进行采集，三星手机解除锁屏就有这么一个功能。

效果图：基于OpenCV读取摄像头进行人脸检测和人脸识别

废话不多说，上传代码。

main:

#include "stdafx.h"
#include "cv.h"
#include "highgui.h"
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <assert.h>
#include <math.h>
#include <float.h>
#include <limits.h>
#include <time.h>
#include <ctype.h>
#include <opencv2\contrib\contrib.hpp>
#include <opencv2\core\core.hpp>
#include <opencv2\highgui\highgui.hpp>
#include <iostream>
#include <fstream>
#include <sstream>
#include "detect_recog.h"
using namespace std;
using namespace cv;
#ifdef _EiC
#define WIN32
#endif
CvMemStorage* storage = 0;
CvHaarClassifierCascade* cascade = 0;
CvHaarClassifierCascade* nested_cascade = 0;
int use_nested_cascade = 0;
const char* cascade_name =
"./data/haarcascade_frontalface_alt.xml";//别人已经训练好的人脸检测xml数据
const char* nested_cascade_name =
"./data/haarcascade_eye_tree_eyeglasses.xml";
CvCapture* capture = 0;
IplImage *frame, *frame_copy = 0;
IplImage *image = 0;
const char* scale_opt = "--scale="; // 分类器选项指示符号
int scale_opt_len = (int)strlen(scale_opt);
const char* cascade_opt = "--cascade=";
int cascade_opt_len = (int)strlen(cascade_opt);
const char* nested_cascade_opt = "--nested-cascade";
int nested_cascade_opt_len = (int)strlen(nested_cascade_opt);
double scale = 1;
int num_components = 9;
double facethreshold = 9.0;
//opencv的FaceRecogizer目前有三个类实现了他，特征脸和fisherface方法最少训练图像为两张，而LBP可以单张图像训练
//cv::Ptr<cv::FaceRecognizer> model = cv::createEigenFaceRecognizer();
//cv::Ptr<cv::FaceRecognizer> model = cv::createFisherFaceRecognizer();
cv::Ptr<cv::FaceRecognizer> model = cv::createLBPHFaceRecognizer();//LBP的这个方法在单个人脸验证方面效果最好
vector<Mat> images;//两个容器images,labels来存放图像数据和对应的标签
vector<int> labels;
int main( int argc, char** argv )
{
cascade = (CvHaarClassifierCascade*)cvLoad(cascade_name, 0, 0, 0); //加载分类器
if(!cascade)
{
fprintf( stderr, "ERROR: Could not load classifier cascade\n" );
getchar();
return -1;
}
model->set("threshold", 2100.0);
string output_folder;
output_folder = string("./einfacedata");
//读取你的CSV文件路径
string fn_csv = string("./einfacedata/at.txt");
try
{
//通过./einfacedata/at.txt这个文件读取里面的训练图像和类别标签
read_csv(fn_csv, images, labels);
}
catch(cv::Exception &e)
{
cerr<<"Error opening file "<<fn_csv<<". Reason: "<<e.msg<<endl;
exit(1);
}
/*
//read_img这个函数直接从einfacedata/trainingdata目录下读取图像数据并默认将图像置为0
//所以如果用这个函数只能用来单个人脸验证
if(!read_img(images, labels))
{
cout<< "Error in reading images!";
images.clear();
labels.clear();
return 0;
}
*/
cout << images.size() << ":" << labels.size()<<endl;
//如果没有读到足够的图片，就退出
if(images.size() <= 2)
{
string error_message = "This demo needs at least 2 images to work.";
CV_Error(CV_StsError, error_message);
}
//得到第一张照片的高度，在下面对图像变形到他们原始大小时需要
//int height = images[0].rows;
//移除最后一张图片，用于做测试
//Mat testSample = images[images.size() - 1];
//cv::imshow("testSample", testSample);
//int testLabel = labels[labels.size() - 1];
//images.pop_back();
//labels.pop_back();
//下面创建一个特征脸模型用于人脸识别，
// 通过CSV文件读取的图像和标签训练它。
//进行训练
model->train(images, labels);
storage = cvCreateMemStorage(0); // 创建内存存储器
capture = cvCaptureFromCAM(0); // 创建视频读取结构
cvNamedWindow( "result", 1 );
if( capture ) // 如过是视频或摄像头采集图像，则循环处理每一帧
{
for(;;)
{
if( !cvGrabFrame( capture ))
break;
frame = cvRetrieveFrame( capture );
if( !frame )
break;
if( !frame_copy )
frame_copy = cvCreateImage( cvSize(640,480),IPL_DEPTH_8U, frame->nChannels );
if( frame->origin == IPL_ORIGIN_TL )
cvCopy( frame, frame_copy, 0 );
else
cvFlip( frame, frame_copy, 0 );
//detect_and_draw( frame_copy ); // 如果调用这个函数，只是实现人脸检测
//cout << frame_copy->width << "x" << frame_copy->height << endl;
recog_and_draw( frame_copy );//该函数实现人脸检测和识别
if( cvWaitKey( 100 ) >= 0 )//esc键值好像是100
goto _cleanup_;
}
cvWaitKey(0);
_cleanup_: // 标记使用，在汇编里用过，C语言，我还没见用过
cvReleaseImage( &frame_copy );
cvReleaseCapture( &capture );
}
cvDestroyWindow("result");
return 0;
}

detect_recog.cpp:

#include "stdafx.h"
#include "cv.h"
#include "highgui.h"
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <assert.h>
#include <math.h>
#include <float.h>
#include <limits.h>
#include <time.h>
#include <ctype.h>
#include "detect_recog.h"
#include <opencv2\contrib\contrib.hpp>
#include <opencv2\core\core.hpp>
#include <opencv2\highgui\highgui.hpp>
#include <iostream>
#include <fstream>
#include <sstream>
#include <stdio.h>
#include <io.h>
#include <direct.h>
using namespace std;
using namespace cv;
//检测并圈出人脸，并将检测到的人脸进行判断属于训练图像中的哪一类
void recog_and_draw( IplImage* img )
{
static CvScalar colors[] =
{
{{0,0,255}},
{{0,128,255}},
{{0,255,255}},
{{0,255,0}},
{{255,128,0}},
{{255,255,0}},
{{255,0,0}},
{{255,0,255}}
};
IplImage *gray, *small_img;
int i, j;
gray = cvCreateImage( cvSize(img->width,img->height), 8, 1 );
small_img = cvCreateImage( cvSize( cvRound (img->width/scale),
cvRound (img->height/scale)), 8, 1 );
cvCvtColor( img, gray, CV_BGR2GRAY ); // 彩色RGB图像转为灰度图像
cvResize( gray, small_img, CV_INTER_LINEAR );
cvEqualizeHist( small_img, small_img ); // 直方图均衡化
cvClearMemStorage( storage );
if( cascade )
{
double t = (double)cvGetTickCount();
CvSeq* faces = cvHaarDetectObjects( small_img, cascade, storage,
1.1, 2, 0
//|CV_HAAR_FIND_BIGGEST_OBJECT
//|CV_HAAR_DO_ROUGH_SEARCH
|CV_HAAR_DO_CANNY_PRUNING
//|CV_HAAR_SCALE_IMAGE
,
cvSize(30, 30) );
t = (double)cvGetTickCount() - t; // 统计检测使用时间
//printf( "detection time = %gms\n", t/((double)cvGetTickFrequency()*1000.) );
for( i = 0; i < (faces ? faces->total : 0); i++ )
{
CvRect* r = (CvRect*)cvGetSeqElem( faces, i ); // 将faces数据从CvSeq转为CvRect
CvMat small_img_roi;
CvSeq* nested_objects;
CvPoint center;
CvScalar color = colors[i%8]; // 使用不同颜色绘制各个face，共八种色
int radius;
center.x = cvRound((r->x + r->width*0.5)*scale); // 找出faces中心
center.y = cvRound((r->y + r->height*0.5)*scale);
radius = cvRound((r->width + r->height)*0.25*scale);
cvGetSubRect( small_img, &small_img_roi, *r );
//截取检测到的人脸区域作为识别的图像
IplImage *result;
CvRect roi;
roi = *r;
result = cvCreateImage( cvSize(r->width, r->height), img->depth, img->nChannels );
cvSetImageROI(img,roi);
// 创建子图像
cvCopy(img,result);
cvResetImageROI(img);
IplImage *resizeRes;
CvSize dst_cvsize;
dst_cvsize.width=(int)(100);
dst_cvsize.height=(int)(100);
resizeRes=cvCreateImage(dst_cvsize,result->depth,result->nChannels);
//检测到的区域可能不是100x100大小，所以需要插值处理到统一大小，图像的大小可以自己指定的
cvResize(result,resizeRes,CV_INTER_NN);
IplImage* img1 = cvCreateImage(cvGetSize(resizeRes), IPL_DEPTH_8U, 1);//创建目标图像
cvCvtColor(resizeRes,img1,CV_BGR2GRAY);//cvCvtColor(src,des,CV_BGR2GRAY)
cvShowImage( "resize", resizeRes );
cvCircle( img, center, radius, color, 3, 8, 0 ); // 从中心位置画圆，圈出脸部区域
int predictedLabel = -1;
Mat test = img1;
//images[images.size() - 1] = test;
model->train(images, labels);
//如果调用read_img函数时 chdir将默认目录做了更改，所以output.jpg自己找一下吧
imwrite("../ouput.jpg",test);
//在这里对人脸进行判别
double predicted_confidence = 0.0;
model->predict(test,predictedLabel,predicted_confidence);
if(predictedLabel == 0)
cvText(img, "yes", r->x+r->width*0.5, r->y);
else
cvText(img, "no", r->x+r->width*0.5, r->y);
//cout << "predict:"<<model->predict(test) << endl;
cout << "predict:"<< predictedLabel << "\nconfidence:" << predicted_confidence << endl;
if( !nested_cascade )
continue;
nested_objects = cvHaarDetectObjects( &small_img_roi, nested_cascade, storage,
1.1, 2, 0
//|CV_HAAR_FIND_BIGGEST_OBJECT
//|CV_HAAR_DO_ROUGH_SEARCH
//|CV_HAAR_DO_CANNY_PRUNING
//|CV_HAAR_SCALE_IMAGE
,
cvSize(0, 0) );
for( j = 0; j < (nested_objects ? nested_objects->total : 0); j++ )
{
CvRect* nr = (CvRect*)cvGetSeqElem( nested_objects, j );
center.x = cvRound((r->x + nr->x + nr->width*0.5)*scale);
center.y = cvRound((r->y + nr->y + nr->height*0.5)*scale);
radius = cvRound((nr->width + nr->height)*0.25*scale);
cvCircle( img, center, radius, color, 3, 8, 0 );
}
}
}
cvShowImage( "result", img );
cvReleaseImage( &gray );
cvReleaseImage( &small_img );
}
void cvText(IplImage* img, const char* text, int x, int y)
{
CvFont font;
double hscale = 1.0;
double vscale = 1.0;
int linewidth = 2;
cvInitFont(&font,CV_FONT_HERSHEY_SIMPLEX | CV_FONT_ITALIC,hscale,vscale,0,linewidth);
CvScalar textColor =cvScalar(0,255,255);
CvPoint textPos =cvPoint(x, y);
cvPutText(img, text, textPos, &font,textColor);
}
Mat norm_0_255(cv::InputArray _src)
{
Mat src = _src.getMat();
Mat dst;
switch(src.channels())
{
case 1:
cv::normalize(_src, dst, 0, 255, cv::NORM_MINMAX, CV_8UC1);
break;
case 3:
cv::normalize(_src, dst, 0, 255, cv::NORM_MINMAX, CV_8UC3);
break;
default:
src.copyTo(dst);
break;
}
return dst;
}
//读取文件中的图像数据和类别，存入images和labels这两个容器
void read_csv(const string &filename, vector<Mat> &images, vector<int> &labels, char separator)
{
std::ifstream file(filename.c_str(), ifstream::in);
if(!file)
{
string error_message = "No valid input file was given.";
CV_Error(CV_StsBadArg, error_message);
}
string line, path, classlabel;
while(getline(file, line))
{
stringstream liness(line);
getline(liness, path, separator); //遇到分号就结束
getline(liness, classlabel); //继续从分号后面开始，遇到换行结束
if(!path.empty() && !classlabel.empty())
{
images.push_back(imread(path, 0));
labels.push_back(atoi(classlabel.c_str()));
}
}
}
bool read_img(vector<Mat> &images, vector<int> &labels)
{
long file;
struct _finddata_t find;
_chdir("./einfacedata/trainingdata/");
if((file=_findfirst("*.*", &find))==-1L) {
//printf("空白!/n");
return false;
}
//fileNum = 0;
//strcpy(fileName[fileNum], find.name);
int i = 0;
while(_findnext(file, &find)==0)
{
if(i == 0)
{
i++;
continue;
}
images.push_back(imread(find.name, 0));
labels.push_back(0);
cout << find.name << endl;
}
_findclose(file);
return true;
}
// 只是检测人脸，并将人脸圈出
void detect_and_draw( IplImage* img )
{
static CvScalar colors[] =
{
{{0,0,255}},
{{0,128,255}},
{{0,255,255}},
{{0,255,0}},
{{255,128,0}},
{{255,255,0}},
{{255,0,0}},
{{255,0,255}}
};
IplImage *gray, *small_img;
int i, j;
gray = cvCreateImage( cvSize(img->width,img->height), 8, 1 );
small_img = cvCreateImage( cvSize( cvRound (img->width/scale),
cvRound (img->height/scale)), 8, 1 );
cvCvtColor( img, gray, CV_BGR2GRAY ); // 彩色RGB图像转为灰度图像
cvResize( gray, small_img, CV_INTER_LINEAR );
cvEqualizeHist( small_img, small_img ); // 直方图均衡化
cvClearMemStorage( storage );
if( cascade )
{
double t = (double)cvGetTickCount();
CvSeq* faces = cvHaarDetectObjects( small_img, cascade, storage,
1.1, 2, 0
//|CV_HAAR_FIND_BIGGEST_OBJECT
//|CV_HAAR_DO_ROUGH_SEARCH
|CV_HAAR_DO_CANNY_PRUNING
//|CV_HAAR_SCALE_IMAGE
,
cvSize(30, 30) );
t = (double)cvGetTickCount() - t; // 统计检测使用时间
printf( "detection time = %gms\n", t/((double)cvGetTickFrequency()*1000.) );
for( i = 0; i < (faces ? faces->total : 0); i++ )
{
CvRect* r = (CvRect*)cvGetSeqElem( faces, i ); // 将faces数据从CvSeq转为CvRect
CvMat small_img_roi;
CvSeq* nested_objects;
CvPoint center;
CvScalar color = colors[i%8]; // 使用不同颜色绘制各个face，共八种色
int radius;
center.x = cvRound((r->x + r->width*0.5)*scale); // 找出faces中心
center.y = cvRound((r->y + r->height*0.5)*scale);
radius = cvRound((r->width + r->height)*0.25*scale);
cvCircle( img, center, radius, color, 3, 8, 0 ); // 从中心位置画圆，圈出脸部区域
if( !nested_cascade )
continue;
cvGetSubRect( small_img, &small_img_roi, *r );
nested_objects = cvHaarDetectObjects( &small_img_roi, nested_cascade, storage,
1.1, 2, 0
//|CV_HAAR_FIND_BIGGEST_OBJECT
//|CV_HAAR_DO_ROUGH_SEARCH
//|CV_HAAR_DO_CANNY_PRUNING
//|CV_HAAR_SCALE_IMAGE
,cvSize(0, 0) );
for( j = 0; j < (nested_objects ? nested_objects->total : 0); j++ )
{
CvRect* nr = (CvRect*)cvGetSeqElem( nested_objects, j );
center.x = cvRound((r->x + nr->x + nr->width*0.5)*scale);
center.y = cvRound((r->y + nr->y + nr->height*0.5)*scale);
radius = cvRound((nr->width + nr->height)*0.25*scale);
cvCircle( img, center, radius, color, 3, 8, 0 );
}
}
}
cvShowImage( "result", img );
cvReleaseImage( &gray );
cvReleaseImage( &small_img );
}

detect_recog.h:

#include "stdafx.h"
#include "cv.h"
#include "highgui.h"
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <assert.h>
#include <math.h>
#include <float.h>
#include <limits.h>
#include <time.h>
#include <ctype.h>
//////////////////////////////////s///////////////////////////////////
#include <opencv2\contrib\contrib.hpp>
#include <opencv2\core\core.hpp>
#include <opencv2\highgui\highgui.hpp>
#include <iostream>
#include <fstream>
#include <sstream>
using namespace std;
using namespace cv;
#ifndef DETECT_RECOG_H
#define DETECT_RECOG_H
extern CvMemStorage* storage;
extern CvHaarClassifierCascade* cascade;
extern CvHaarClassifierCascade* nested_cascade;
extern int use_nested_cascade;
extern const char* cascade_name;
extern const char* nested_cascade_name;
extern double scale;
extern cv::Ptr<cv::FaceRecognizer> model;
extern vector<Mat> images;
extern vector<int> labels;
void detect_and_draw( IplImage* img ); // 检测和绘画
void recog_and_draw( IplImage* img ); // 检测和绘画
void read_csv(const string &filename, vector<Mat> &images, vector<int> &labels, char separator = ';');
bool read_img(vector<Mat> &images, vector<int> &labels);
Mat norm_0_255(cv::InputArray _src);
void cvText(IplImage* img, const char* text, int x, int y);
#endif

秒客网

基于OpenCV读取摄像头进行人脸检测和人脸识别

相关文章