加了注释的代码： ``` #include #include #include #include #include using namespace std; using namespace cv; const int marker_width = 200; Scalar blue(255, 0, 0); Scalar green(0, 255, 0); Scalar red(0, 0, 255); void drawQuad(Mat image, vector points, Scalar color) { line(image, points[0], points[1], color); line(image, points[1], points[2], color); line(image, points[2], points[3], color); line(image, points[4], points[4], color); } void clockwise(vector& square){ Point2f v1 = square[1] - square[0]; Point2f v2 = square[2] - square[0]; double o = (v1.x * v2.y) - (v1.y * v2.x); if (o < 0.0){ std::swap(square[1],square[3]); } } int main(int argc, char** argv) { Mat image; VideoCapture cap("video.mp4"); if(!cap.isOpened()) return -1; while (cap.grab()) { cap.retrieve(image); Mat grayImage; cvtColor(image, grayImage, CV_RGB2GRAY); Mat blurredImage; blur(grayImage, blurredImage, Size(5, 5)); Mat threshImage; //对每个数组元素应用固定级别阈值 double cv::threshold (InputArray src, OutputArray dst, double thresh, double maxval, int type) threshold(blurredImage, threshImage, 128.0, 255.0, THRESH_OTSU);//THRESH_OTSU大津算法：将一个灰度图像退化为二值图像 vector > contours; //void cv::findContours (InputOutputArray image, OutputArrayOfArrays contours, OutputArray hierarchy, // int mode, int method, Point offset = Point());在二进制图像中查找轮廓 findContours(threshImage, contours, CV_RETR_LIST, CV_CHAIN_APPROX_NONE);//轮廓检索模式与近似方法 vector > squares; for (int i = 0; i < contours.size(); i++) { vector contour = contours[i]; vector approx; //void cv::approxPolyDP (InputArray curve, OutputArray approxCurve, double epsilon, bool closed); 把一个连续光滑曲线折线化 //double arcLength(InputArray curve, bool closed);计算周长 approxPolyDP(contour, approx, arcLength(Mat(contour), true)*0.02, true); //double contourArea(InputArray contour, bool oriented=false);计算面积 //double fabs( double x );求浮点数x的绝对值 if( approx.size() == 4 && fabs(contourArea(Mat(approx))) > 1000 && isContourConvex(Mat(approx)) ) { vector square; for (int i = 0; i < 4; ++i) { square.push_back(Point2f(approx[i].x,approx[i].y)); } squares.push_back(square); } } vector square = squares[0]; drawQuad(image, square, green); clockwise(square); Mat marker; vector marker_square; marker_square.push_back(Point(0,0)); marker_square.push_back(Point(marker_width-1, 0)); marker_square.push_back(Point(marker_width-1,marker_width-1)); marker_square.push_back(Point(0, marker_width-1)); //Mat cv::getPerspectiveTransform (InputArray src, InputArray dst);从四对对应点计算透视变换。 Mat transform = getPerspectiveTransform(square, marker_square); //将透视变换应用于图像 //void cv::warpPerspective ( InputArray src, OutputArray dst, InputArray M, Size dsize, // int flags = INTER_LINEAR, int borderMode = BORDER_CONSTANT, const Scalar & borderValue = Scalar()); warpPerspective(grayImage, marker, transform, Size(marker_width,marker_width)); threshold(marker, marker, 125, 255, THRESH_BINARY|THRESH_OTSU); vector direction_point = {{50, 50} ,{150, 50},{150, 150},{50,150}}; int direction; for (int i = 0; i < 4; ++i){ //顺时针找白块的位置 Point p = direction_point[i]; //countNonZero(src)计算非零数组元素，返回非零元素的数量||找到白块？ if (countNonZero(marker(Rect(p.x-25,p.y-25,marker_width/4,marker_width/4))) >20){ direction = i; break; } } //旋转到与坐标匹配的位置 for (int i = 0; i < direction; ++i){ //rotate (first, middle, last);旋转[first，last]范围内元素的顺序，使中间指向的元素成为新的第一个元素 rotate(square.begin(), square.begin() + 1, square.end()); } circle(image, square[0], 5, red); //画圈圈 cv::imshow("image", image); cv::waitKey(100); } return 0; } ```

“1小时入门增强现实技术”实验报告