mitkHandeyeSensitivityTest.cxx File Reference

Include dependency graph for mitkHandeyeSensitivityTest.cxx:

Functions
cv::Mat	WorldToLens (cv::Mat PointInWorldCoordinates, cv::Mat TrackerToWorld, cv::Mat TrackerToLens)
cv::Mat	LensToWorld (cv::Mat PointInLensCoordinates, cv::Mat TrackerToWorld, cv::Mat TrackerToLens)
int	mitkHandeyeSensitivityTest (int argc, char *argv[])

Function Documentation

cv::Mat LensToWorld	(	cv::Mat	PointInLensCoordinates,
		cv::Mat	TrackerToWorld,
		cv::Mat	TrackerToLens
	)

int mitkHandeyeSensitivityTest	(	int	argc,
		char *	argv[]
	)

cv::Mat WorldToLens	(	cv::Mat	PointInWorldCoordinates,
		cv::Mat	TrackerToWorld,
		cv::Mat	TrackerToLens
	)

This test attempts to answer the question of how important errors in the hand eye calibration are. Start with a point in world space. p We estimate it's position (p est) in world space using it's measured position in screen coordinates and T screen to world. Which is TScreenToLens (intrinsic) * TLensToTracker (handeye) * TTrackerToWorld Shouldn't need intrinsic, we estimate it's position in world space using it's measured position in lens coordinates. The estimated position may be very wrong due to the handeye error, but we are only interested in the sunsequent error in it's position relative to the lens. so we now have two p1, (relative to the lens) the actual and the estimated. p1 actual = p * WorldToTracker * handeye-1 and p1 (est) = p (est) * WorldToTracker * handeye (est) -1 and error = p1 actual - p1 (est) - or vice versa

The purpose of this program is to determine the error for a set of increasingly erroreous handeyes, to try and show how handeye is related to projection error in this case