#include <face/face_recognition.h>
void preDeal_original_img(const std::string recognitionPath, const std::string cascadeName)
{
std::ifstream fin;
fin.open(recognitionPath);
if (!fin)
{
std::cout << "Cannot open " + recognitionPath << std::endl;
return;
}
// --Detection
cv::CascadeClassifier cascade;
cascade.load(cascadeName);
if (cascade.empty())
{
std::cout << "Cascade path error!" << std::endl;
return;
}
double scale = 1.3;
std::vector<cv::Rect> faces;
cv::Mat gray;
std::string name;
std::string camera_face = "../camera_face/";
while (getline(fin, name)) {
if (name.empty())
{
continue;
}
name.erase(0, name.find_first_not_of(" \t"));
name.erase(name.find_last_not_of(" \t") + 1);
// Read Image
cv::Mat img = cv::imread(name);
if (img.empty())
{
continue;
}
cv::Mat_<uchar> image;
if (img.channels() != 1)
{
cvtColor(img, image, CV_BGR2GRAY);
image.convertTo(image, CV_8UC1);
}
else {
image = img;
}
// 改变图像
cv::Mat smallImg(cvRound(image.rows / scale), cvRound(image.cols / scale), CV_8UC1); //cvRound对double型数据进行四舍五入
cv::resize(image, smallImg, smallImg.size(), 0, 0, cv::INTER_LINEAR);
cv::equalizeHist(smallImg, smallImg); //equalizeHist提高图像的亮度和对比度
// --Detection
cascade.detectMultiScale(smallImg, faces,
1.1, 3, 0
/*|CV_HAAR_FIND_BIGGEST_OBJECT
|CV_HAAR_DO_ROUGH_SEARCH*/
| CV_HAAR_SCALE_IMAGE
,
cv::Size(30, 30));
if (faces.size() > 0)
{
size_t pos = name.find_last_of('\\');
std::string filename = name.substr(pos + 1);
if (-1 == _access(camera_face.c_str(), 0))
{
_mkdir(camera_face.c_str());
}
filename = camera_face + filename;
std::cout << filename << std::endl;
cv::imwrite(filename, img);
}
}
fin.close();
//处理后的图片路径名写入Path_Image.txt中
std::string getImgPathTxt = "cd " + camera_face + " && dir /b/s/p/w *.jpg > Path_Images.txt";
system(getImgPathTxt.c_str());
}
std::vector<std::pair<cv::Mat, std::string >> get_CropFace_And_ImgPathName(const std::string recognitionPath, const std::string cascadeName)
{
std::vector<std::pair<cv::Mat, std::string>> cropFaceAndImgPathNames;
std::pair<cv::Mat, std::string> cropFaceAndImgPathName;
cv::CascadeClassifier cascade;
cascade.load(cascadeName);
if (cascade.empty())
{
std::cout << "Cascade path error!" << std::endl;
return std::vector<std::pair<cv::Mat, std::string >>();
}
std::ifstream fdatabase;
fdatabase.open(recognitionPath);
if (!fdatabase)
{
std::cout << "Cannot open " + recognitionPath << std::endl;
return std::vector<std::pair<cv::Mat, std::string >>();
}
double scale = 1.3;
std::vector<cv::Rect> faces;
cv::Mat gray;
std::string name;
std::cout << "[";
while (getline(fdatabase, name)) {
if (name.empty())
{
continue;
}
name.erase(0, name.find_first_not_of(" \t"));
name.erase(name.find_last_not_of(" \t") + 1);
// Read Image
cv::Mat img = cv::imread(name);
if (img.empty())
{
continue;
}
cv::Mat image;
if (img.channels() != 1)
{
cvtColor(img, image, CV_BGR2GRAY);
image.convertTo(image, CV_8UC1);
}
else {
image = img;
}
// Read Opencv Detection Bbx
cv::Mat smallImg(cvRound(image.rows / scale), cvRound(image.cols / scale), CV_8UC1); //cvRound对double型数据进行四舍五入
cv::resize(image, smallImg, smallImg.size(), 0, 0, cv::INTER_LINEAR);
cv::equalizeHist(smallImg, smallImg); //equalizeHist提高图像的亮度和对比度
// --Detection
cascade.detectMultiScale(smallImg, faces,
1.1, 3, 0
/*|CV_HAAR_FIND_BIGGEST_OBJECT
|CV_HAAR_DO_ROUGH_SEARCH*/
| CV_HAAR_SCALE_IMAGE
,
cv::Size(30, 30));
for (std::vector<cv::Rect>::iterator r = faces.begin(); r != faces.end(); r++)
{
cv::Rect face;
face.x = int(r->x * scale);
face.y = int(r->y * scale);
face.width = int(r->width * scale);
face.height = int(r->height * scale);
// 边界检查，左边界，上边界，右边界，下边界。
/*face.x = face.x < 1 ? 1 : face.x;
face.y = face.y < 1 ? 1 : face.y;
face.width = (face.x + face.width) > image.cols ? (image.cols - face.x) : face.width;
face.height = (face.y + face.height) > image.rows ? (image.rows - face.y) : face.height;*/
cv::Mat cropFace;
cropFace = img(face);
/*cv::moveWindow("cropface", 960 - cropFace.cols / 2, 540 - cropFace.rows / 2);
cv::imshow("cropface", cropFace);
cv::waitKey(100);
cv::destroyWindow("cropface");*/
cropFaceAndImgPathName = make_pair(cropFace, name); //cropFaceAndImgPathName = std::pair<cv::Mat, std::string>(cropFace, name);
cropFaceAndImgPathNames.push_back(cropFaceAndImgPathName);
std::cout << '.';
}
}
fdatabase.close();
std::cout << "]" << std::endl;
return cropFaceAndImgPathNames;
}
bool matchFace(cv::Mat detectFace, cv::Mat dbFace)
{
IplImage* srcImg = cvCloneImage(&(IplImage)detectFace);
IplImage* dstImg = cvCloneImage(&(IplImage)dbFace);
IplImage* src;
IplImage* dst;
if (srcImg->nChannels != 1)
{
src = cvCreateImage(cvSize(srcImg->width, srcImg->height), srcImg->depth, 1);
cvCvtColor(srcImg, src, CV_BGR2GRAY);
}
if (dstImg->nChannels != 1)
{
dst = cvCreateImage(cvSize(dstImg->width, dstImg->height), dstImg->depth, 1);
cvCvtColor(dstImg, dst, CV_BGR2GRAY);
}
int histogramBins = 256;
float histogramRange1[2] = { 0, 255 };
float* histogramRange[1] = { &histogramRange1[0] };
CvHistogram* Histogram1 = cvCreateHist(1, &histogramBins, CV_HIST_ARRAY, histogramRange);
CvHistogram* Histogram2 = cvCreateHist(1, &histogramBins, CV_HIST_ARRAY, histogramRange);
cvCalcHist(&src, Histogram1);
cvCalcHist(&dst, Histogram2);
cvNormalizeHist(Histogram1, 1);
cvNormalizeHist(Histogram2, 1);
// CV_COMP_CHISQR,CV_COMP_BHATTACHARYYA这两种都可以用来做直方图的比较，值越小，说明图形越相似
//printf("CV_COMP_CHISQR : %.4f\n", cvCompareHist(Histogram1, Histogram2, CV_COMP_CHISQR));
//printf("CV_COMP_BHATTACHARYYA : %.4f\n", cvCompareHist(Histogram1, Histogram2, CV_COMP_BHATTACHARYYA));
// CV_COMP_CORREL, CV_COMP_INTERSECT这两种直方图的比较，值越大，说明图形越相似
//printf("CV_COMP_CORREL : %.4f\n", cvCompareHist(Histogram1, Histogram2, CV_COMP_CORREL));
//printf("CV_COMP_INTERSECT : %.4f\n", cvCompareHist(Histogram1, Histogram2, CV_COMP_INTERSECT));
double simility = cvCompareHist(Histogram1, Histogram2, CV_COMP_CHISQR);
if (simility > 0.5)
{
return false;
}
return true;
}
void face_recognition(std::string recognitionPath, const std::string cascadeName)
{
bool isPreDeal = false;
if (isPreDeal) //是否进行预处理
{
preDeal_original_img(recognitionPath, cascadeName);
recognitionPath = "../camera_face/";
}
//获取数据库中人脸图像
std::string face_Database = "../face_database/Path_Images.txt";
std::vector<std::pair<cv::Mat, std::string>> cropFaceAndImgPathNames;
std::cout << "开始数据库中人脸数据的读取..." << std::endl;
cropFaceAndImgPathNames = get_CropFace_And_ImgPathName(face_Database, cascadeName);
std::cout << "结束数据库中人脸数据的读取。" << std::endl;
//开始人脸匹配
std::ifstream frecognition;
frecognition.open(recognitionPath);
if (!frecognition)
{
std::cout << "Images path Error!" << std::endl;
return;
}
cv::CascadeClassifier cascade;
cascade.load(cascadeName);
if (cascade.empty())
{
std::cout << "Cascade path error!" << std::endl;
return;
}
double scale = 1.3;
std::vector<cv::Rect> faces;
cv::Mat gray;
std::string name;
bool isExist = false; //数据库中是否存在该匹配文件
while (getline(frecognition, name)) {
if (name.empty())
{
continue;
}
name.erase(0, name.find_first_not_of(" \t"));
name.erase(name.find_last_not_of(" \t") + 1);
// Read Image
cv::Mat img = cv::imread(name);
cv::Mat image;
if (img.channels() != 1)
{
cvtColor(img, image, CV_BGR2GRAY);
image.convertTo(image, CV_8UC1);
}
else {
image = img;
}
// Read Opencv Detection Bbx
cv::Mat smallImg(cvRound(image.rows / scale), cvRound(image.cols / scale), CV_8UC1); //cvRound对double型数据进行四舍五入
cv::resize(image, smallImg, smallImg.size(), 0, 0, cv::INTER_LINEAR);
cv::equalizeHist(smallImg, smallImg); //equalizeHist提高图像的亮度和对比度
// --Detection
cascade.detectMultiScale(smallImg, faces,
1.1, 3, 0
/*|CV_HAAR_FIND_BIGGEST_OBJECT
|CV_HAAR_DO_ROUGH_SEARCH*/
| CV_HAAR_SCALE_IMAGE
,
cv::Size(30, 30));
for (std::vector<cv::Rect>::iterator r = faces.begin(); r != faces.end(); r++)
{
cv::Rect face;
face.x = int(r->x * scale);
face.y = int(r->y * scale);
face.width = int(r->width * scale);
face.height = int(r->height * scale);
cv::Mat detectFace = img(face);
for (std::vector<std::pair<cv::Mat, std::string>>::iterator dbFace = cropFaceAndImgPathNames.begin(); dbFace != cropFaceAndImgPathNames.end(); dbFace++)
{
std::pair<cv::Mat, std::string> dbFaceImg = *dbFace;
bool isMatch = matchFace(detectFace, dbFaceImg.first);
if (isMatch) {
std::cout << name + " Matching " + dbFaceImg.second + " successful!" << std::endl;
cv::imshow("detectFace", detectFace);
cv::imshow("databaseFace", dbFaceImg.first);
cv::waitKey(200);
cv::destroyWindow("detectFace");
cv::destroyWindow("databaseFace");
isExist = true;
}
}
}
}
if (!isExist)
{
std::cout << name + " Matching failed!" << std::endl;
}
frecognition.close();
}
192行，193行出现错误，请问这种错误应该怎么改啊
```严重性 代码 说明 项目 文件 行 禁止显示状态
错误(活动) E0312 不存在用户定义的从 "cv::Mat" 到 "IplImage" 的适当转换 ConsoleApplication1 D:\Clion\demo\ConsoleApplication1\face_recognition.cpp 192
换成 IplImage* srcImg = detectFace.clone()看下，应该是后面的那个强制转换失败了，或者 IplImage* srcImg = cvCloneImage((IplImage*)detectFace);