industrial_calibration::ExtrinsicMultiStaticCameraOnlyResult Struct Reference

Public Attributes
bool	converged
	Whether the underlying solver converged. If this is false, your calibration did not go well. If this is true, your calibration MAY have gone well.
double	initial_cost_per_obs
	The initial reprojection error (in pixels) per residual based on your input guesses.
double	final_cost_per_obs
	The final reprojection error (in pixels) per residual after optimization. Note that each circle has two residuals: a U and V error in the image. So a value of 1.2 means that each circle was described to within 1.2 pixels in X and 1.2 pixels in Y.
std::vector< Eigen::Isometry3d >	base_to_target
	The final calibrated result of "base frame" to "target frame".
std::vector< Eigen::Isometry3d >	base_to_camera
	The final calibrated result of "base frame" to "camera optical frame".
CovarianceResult	covariance

Member Data Documentation

final_cost_per_obs

double industrial_calibration::ExtrinsicMultiStaticCameraOnlyResult::final_cost_per_obs

A low value here is encouraging if you had a diversity of images. If you took few images, you can get a low score without getting a calibration that describes your workcell.