# `OpenCVFaceDetection` `OpenCV`人脸检测`C++`项目`Demo` + `LBP `算法
[TOC]
## 一、参考
[OpenCV 人脸样本训练](https://www.jianshu.com/p/c3120e8301aa)
[OpenCV 中的 Mat](https://www.jianshu.com/p/c0588be86ed6)
[级联分类器训练](http://www.opencv.org.cn/opencvdoc/2.3.2/html/doc/user_guide/ug_traincascade.html)
[LBP特征](https://www.jianshu.com/p/47ec293bf55d)
## 二、实操
### 2.1 `CMakeLists.txt`
```cmake
cmake_minimum_required (VERSION 3.8)
# 导入opencv头文件
include_directories("D:/CommonDev/opencv/opencv4.1.2/opencv/build/include")
# 导入opencv动态库路径
link_directories("D:/CommonDev/opencv/opencv4.1.2/opencv/build/x64/vc15/lib")
# 将源代码添加到此项目的可执行文件。
add_executable (OpenCVFaceDetection "OpenCVFaceDetection.cpp" "OpenCVFaceDetection.h")
# 将名字为“opencv_world412d”的动态库预编译到“NeOpencvStructure”可执行文件
target_link_libraries(OpenCVFaceDetection opencv_world412d)
```
### 2.2 `OpenCVFaceDetection.h`
```c++
#pragma once
#include <iostream>
#include "opencv2/opencv.hpp"
using namespace std;
using namespace cv;
//#define DETECT //静态检测开关
//#define DYNAMIC_DETECT //动态检测开关
//#define COLECT_SAMPLES //采集人脸正样本
#define LBP_CALCULATE //计算lbp
class CascadeDetectorAdapter : public DetectionBasedTracker::IDetector
{
public:
CascadeDetectorAdapter(cv::Ptr<cv::CascadeClassifier> detector) :
IDetector(),
Detector(detector)
{
CV_Assert(detector);
}
void detect(const cv::Mat& Image, std::vector<cv::Rect>& objects)
{
Detector->detectMultiScale(Image, objects, scaleFactor, minNeighbours, 0, minObjSize, maxObjSize);
}
virtual ~CascadeDetectorAdapter()
{
}
private:
CascadeDetectorAdapter();
cv::Ptr<cv::CascadeClassifier> Detector;
};
```
### 2.3 `OpenCVFaceDetection.cpp`
```c++
#include "OpenCVFaceDetection.h"
void calcLBP(Mat, Mat&);
CascadeClassifier face_cascade_classifier;
cv::Ptr<DetectionBasedTracker> tracker;
int main()
{
//Mat ones = Mat::ones(2, 2, CV_8UC3);
//Mat ones = Mat::eye(3, 3, CV_8UC3);;
//cout << ones << endl;
#ifdef DETECT
if (!face_cascade_classifier.load("D:/CommonDev/opencv/opencv4.1.2/opencv/build/etc/haarcascades/haarcascade_frontalface_alt.xml")) {
cout << "级联分类器加载失败！" << endl;
return -1;
}
#elif DYNAMIC_DETECT
//主检测适配器
cv::Ptr<CascadeDetectorAdapter> mainDetector = makePtr<CascadeDetectorAdapter>(
makePtr<CascadeClassifier>("D:/CommonDev/opencv/opencv4.1.2/opencv/build/etc/haarcascades/haarcascade_frontalface_alt.xml"));
//跟踪检测适配器
cv::Ptr<CascadeDetectorAdapter> trackingDetector = makePtr<CascadeDetectorAdapter>(
makePtr<CascadeClassifier>("D:/CommonDev/opencv/opencv4.1.2/opencv/build/etc/haarcascades/haarcascade_frontalface_alt.xml"));
//创建跟踪器
DetectionBasedTracker::Parameters DetectorParams;
tracker = makePtr<DetectionBasedTracker>(mainDetector, trackingDetector, DetectorParams);
//启动跟踪器
tracker->run();
#endif // DETECT
VideoCapture capture;
capture.open(0);
if (!capture.isOpened()) {
cout << "摄像头打开失败！" << endl;
return -1;
}
Mat frame; //摄像头的图像
Mat gray0, gray;
while (true) {
capture >> frame;
if (frame.empty()) {
cout << "摄像头采集图像失败！" << endl;
return -1;
}
imshow("摄像头", frame); //彩色图
//预处理
cvtColor(frame, gray0, COLOR_RGB2GRAY); //转成灰度图， opencv中：bgr
imshow("黑白相机", gray0);
//equalizeHist(gray0, gray); //直方均衡化，增强对比度
vector<Rect> faces;
#ifdef LBP_CALCULATE
Mat lbp = Mat(gray0.rows - 2, gray0.cols - 2, CV_8UC1);
calcLBP(gray0, lbp);
imshow("lbp图", lbp);
#else
#ifdef DETECT
face_cascade_classifier.detectMultiScale(gray, faces);
#elif DYNAMIC_DETECT
//人脸动态检测处理
tracker->process(gray);
tracker->getObjects(faces);
#endif // DETECT
int i = 0;
for each (Rect face in faces) { //VS特有的语法
rectangle(frame, face, Scalar(0, 0, 255));
#ifdef COLECT_SAMPLES //采集样本
Mat sample;
frame(face).copyTo(sample); //抠图（在彩图frame 上抠图）
resize(sample, sample, Size(24, 24)); //归一化样本大小
cvtColor(sample, sample, COLOR_BGR2GRAY); //灰度化处理
char p[100];
sprintf(p, "C:/Users/Administrator/Desktop/opencv/train/face/pos/%d.jpg", i++);
imwrite(p, sample); //保存样本图像，与imread() 读对应
#endif // COLECT_SAMPLES //采集样本
}
imshow("人脸检测", frame);
#endif //LBP_CALCULATE
//等待30ms
if (waitKey(30) == 27) { //ESC键
break;
}
}
return 0;
}
// 计算LBP图谱算法
// src灰度图
// Mat(高，宽)
void calcLBP(Mat src, Mat& dst) {
for (int i = 1; i < src.rows - 1; i++) //高
{
for (int j = 1; j < src.cols - 1; j++)
{
uchar lbp = 0;
//取（i,j）位置的像素
uchar center = src.at<uchar>(i, j);
//左上角 顺时针
if (src.at<uchar>(i - 1, j - 1) > center) {
lbp += 1 << 7;
}
//正上方
if (src.at<uchar>(i - 1, j) > center) {
lbp += 1 << 6;
}
//右上方
if (src.at<uchar>(i - 1, j + 1) > center) {
lbp += 1 << 5;
}
//正右方
if (src.at<uchar>(i, j + 1) > center) {
lbp += 1 << 4;
}
//右下角
if (src.at<uchar>(i + 1, j + 1) > center) {
lbp += 1 << 3;
}
//正下方
if (src.at<uchar>(i + 1, j) > center) {
lbp += 1 << 2;
}
//左下方
if (src.at<uchar>(i + 1, j - 1) > center) {
lbp += 1 << 1;
}
//左
if (src.at<uchar>(i, j - 1) > center) {
lbp += 1 << 0;
}
dst.at<uchar>(i - 1, j - 1) = lbp;
}
}
}
```
