数字图像处理项目：背景建模跟踪目标（提取轮廓显示矩形框）

#include <opencv2/opencv.hpp>
#include <iostream>

using namespace cv;
using namespace std;

//针处理函数
void processFrame(Mat& binary, vector<Rect>& rect);

int main(int argc, char**) {
	VideoCapture capture;
	capture.open("video_003.avi");
	if (!capture.isOpened()) {
		printf("could not find the video file...\n");
		return -1;
	}
	int frame_count = capture.get(CAP_PROP_FRAME_COUNT);
	int fps = capture.get(CAP_PROP_FPS);
	cout << frame_count << endl;
	cout << fps << endl;
	// create windows
	Mat frame;
	Mat bsmaskMOG2, bsmaskKNN;
	namedWindow("input video", WINDOW_AUTOSIZE);
	namedWindow("MOG2", WINDOW_AUTOSIZE);
	namedWindow("KNN Model", WINDOW_AUTOSIZE);
	Mat kernel = getStructuringElement(MORPH_RECT, Size(3, 3), Point(-1, -1));
	Mat kernel2 = getStructuringElement(MORPH_ELLIPSE, Size(3, 3), Point(-1, -1));
	// intialization BS
	Ptr<BackgroundSubtractor> pMOG2 = createBackgroundSubtractorMOG2();
	/*
	Ptr<backgroundsubtractormog2>
	cv::createBackgroundSubtractorMOG2  (
						int     history = 500,  //用于训练背景的帧数，默认为500帧，如果不手动设置learningRate
						，history就被用于计算当前的learningRate，此时history越大，learningRate越小，背景更新越慢；
						double  varThreshold = 16,  //方差阈值，用于判断当前像素是前景还是背景。一般默认16
						，如果光照变化明显，如阳光下的水面，建议设为25,36，具体去试一下也不是很麻烦，值越大，灵敏度越低；
						bool    detectShadows = true   //是否检测影子，设为true为检测，false为不检测，
						检测影子会增加程序时间复杂度，如无特殊要求，建议设为false；
)
	*/
	Ptr<BackgroundSubtractor> pKNN = createBackgroundSubtractorKNN();
	while (capture.read(frame)) {
		vector<Rect> rect;
		
		// MOG BS
		pMOG2->apply(frame, bsmaskMOG2);
		/*
		virtual void cv::BackgroundSubtractor::apply    (
				InputArray  image,		//image 源图
				OutputArray     fgmask,  //fmask 前景（二值图像）
				double  learningRate = -1   //learningRate 学习速率，
				，值为0-1,为0时背景不更新，为1时逐帧更新，默认为-1，即算法自动更新
				)
		*/
		morphologyEx(bsmaskMOG2, bsmaskMOG2, MORPH_OPEN, kernel, Point(-1, -1));
		//dilate(bsmaskMOG2, bsmaskMOG2, kernel2, Point(-1, -1), 2);// 膨胀
		
	
		imshow("MOG2", bsmaskMOG2);

		// KNN BS mask
		// 最后一个参数 (0-1之间表示学习背景模型时长，-1表示opencv自动设置学习时长
		//，0表示不更改背景模型，1表示背景模型按照最后一帧重新初始化)
		pKNN->apply(frame, bsmaskKNN, -1);
		imshow("KNN Model", bsmaskKNN);
		morphologyEx(bsmaskKNN, bsmaskKNN, MORPH_OPEN, kernel2, Point(-1, -1));
		dilate(bsmaskKNN, bsmaskKNN, kernel2, Point(-1, -1), 4);// 膨胀
		//处理bsmaskKNN二值化图像
		processFrame(bsmaskMOG2, rect);
		for (int j = 0; j < rect.size(); j++)
		{
			rectangle(frame, rect[j], Scalar(0, 0, 255), 2, 8);
		}
		imshow("input video", frame);
		char c = waitKey(100);
		if (c == 27) {
			break;
		}
	}
	capture.release();
	waitKey(0);
	return 0;
}

void processFrame(Mat& binary, vector<Rect>& rect) {
	vector<vector<Point>> contours;
	vector<Vec4i> hireachy;
	findContours(binary, contours, hireachy, RETR_EXTERNAL, CHAIN_APPROX_SIMPLE, Point(0, 0));
	int num = contours.size();
	for (int i = 0; i < num; i++)
	{
		int num_points = contours[i].size();
		if (num_points > 20)
		{
			Rect point_rect = boundingRect(contours[i]);
			rect.push_back(point_rect);
		}
	}

}