Trying to get points while detecting a box

edit

Step 1: I'm trying to detect the box volume. So I first detected the corners. The user will have a snap point and axis, so I will have a predefined point in world coordinates.

Step 2: I have converted the corner points into 3D Positions using the inverse camera internsics. Step 3: I have project rays, I have the ground plane at 0,0,0, those rays will intersect with a plane with normal UP.

Question: Now I want to know what are the points that are on the ground and what are the points that are above the ground. Here is the code

struct Ray
{
  cv::Vec3f origin;
  cv::Vec3f direction;

};

![image description](/upfiles/1508262876513037.png)

bool linePlaneIntersection( Vec3f& contact, Vec3f direction, Vec3f rayOrigin,
  Vec3f normal, Vec3f coord ) {
  // get d value
  float d = normal.dot(coord);

  if ( normal.dot(direction) == 0 ) {
    return false; // No intersection, the line is parallel to the plane
  }

  // Compute the X value for the directed line ray intersecting the plane
  float x = (d - normal.dot(rayOrigin)) / normal.dot(direction);

  // output contact point
  contact = rayOrigin + normalize(direction)*x; //Make sure your ray vector is normalized
  return true;
}

int main( int argc, char** argv )
{


  Mat src, src_copy, edges, dst;
  src = imread( "freezeFrame__1508152029892.png", 0 );

  src_copy = src.clone();



  std::vector< cv::Point2f > corners;

  // maxCorners – The maximum number of corners to return. If there are more corners
  // than that will be found, the strongest of them will be returned
  int maxCorners = 10;

  // qualityLevel – Characterizes the minimal accepted quality of image corners;
  // the value of the parameter is multiplied by the by the best corner quality
  // measure (which is the min eigenvalue, see cornerMinEigenVal() ,
  // or the Harris function response, see cornerHarris() ).
  // The corners, which quality measure is less than the product, will be rejected.
  // For example, if the best corner has the quality measure = 1500,
  // and the qualityLevel=0.01 , then all the corners which quality measure is
  // less than 15 will be rejected.
  double qualityLevel = 0.01;

  // minDistance – The minimum possible Euclidean distance between the returned corners
  double minDistance = 20.;

  // mask – The optional region of interest. If the image is not empty (then it
  // needs to have the type CV_8UC1 and the same size as image ), it will specify
  // the region in which the corners are detected
  cv::Mat mask;

  // blockSize – Size of the averaging block for computing derivative covariation
  // matrix over each pixel neighborhood, see cornerEigenValsAndVecs()
  int blockSize = 3;

  // useHarrisDetector – Indicates, whether to use operator or cornerMinEigenVal()
  bool useHarrisDetector = false;

  // k – Free parameter of Harris detector
  double k = 0.04;

  cv::goodFeaturesToTrack( src, corners, maxCorners, qualityLevel, minDistance, mask, blockSize, useHarrisDetector, k );


  for ( size_t i = 0; i < corners.size(); i++ )
  {
    cv::circle( src, corners[i], 10, cv::Scalar( 0, 255, 255 ), -1 );
  }

  for ( size_t i = 0; i < corners.size(); i++ )
  {
    cv::Point2f pt1 = corners[i];
    for ( int j = 0; j < corners.size(); j++)
    {
      cv::Point2f pt2 = corners[j];
      line( src, pt1, pt2, cv::Scalar( 0, 255, 0 ) );
    }
  }


  Mat cameraIntrinsics( 3, 3, CV_32F );

  cameraIntrinsics.at<float> ...