Question

我正在编写一种算法来稳定视频，我目前的流程是：

使用SURF获取功能
构建功能Kd-Tree然后匹配每两个连续的帧。我正在使用来自openCV的Knnmatch，但不确定是否使用Kd树。
cvFindHomography计算单应性
与单应性矩阵的倒数一起翘曲

现在的代码如下：

void Stab::process()
{
  Mat gray;
  Mat img_keypoints_1;

  cvtColor(frame, gray, COLOR_BGR2GRAY);
  if(first == false)
  {
    prev_frame = frame.clone();
    first = true;
  }

  //-- Step 1: Detect the keypoints using SURF Detector, compute the descriptors
  int minHessian = 400;
  Ptr<SURF> detector = SURF::create( minHessian );
  std::vector<KeyPoint> keypoints1, keypoints2;
  Mat descriptors1, descriptors2;
  detector->detectAndCompute( prev_frame, noArray(), keypoints1, descriptors1 );
  detector->detectAndCompute( frame, noArray(), keypoints2, descriptors2 );


  //-- Step 2: Matching descriptor vectors with a FLANN based matcher
  // Since SURF is a floating-point descriptor NORM_L2 is used
  Ptr<DescriptorMatcher> matcher = DescriptorMatcher::create(DescriptorMatcher::FLANNBASED);
  std::vector< std::vector<DMatch> > knn_matches;
  matcher->knnMatch( descriptors1, descriptors2, knn_matches, 2 );


  //-- Filter matches using the Lowe's ratio test
  const float ratio_thresh = 0.7f;
  std::vector<DMatch> good_matches;
  for (size_t i = 0; i < knn_matches.size(); i++)
  {
      if (knn_matches[i][0].distance < ratio_thresh * knn_matches[i][1].distance)
      {
          good_matches.push_back(knn_matches[i][0]);
      }
  }
    Mat HInv;
  //-- Localize the object
    std::vector<Point2f> obj;
    std::vector<Point2f> scene;
    for( size_t i = 0; i < good_matches.size(); i++ )
    {
        //-- Get the keypoints from the good matches
        obj.push_back( keypoints1[ good_matches[i].queryIdx ].pt );
        scene.push_back( keypoints2[ good_matches[i].trainIdx ].pt );
    }
    Mat H = findHomography( obj, scene, RANSAC );
    HInv = H.inv();

    warpPerspective(frame, im_out, HInv, prev_frame.size());
  prev_frame = frame.clone();
  //waitKey(30);
}

但是，我得到的结果非常糟糕，是否可以进行任何优化，还是应该切换到其他方法？

OpenCV中的视频稳定

0 个答案: