lsst.jointcal
master-g59ec7209dd
|
Class that handles the astrometric least squares problem. More...
#include <AstrometryFit.h>
Public Member Functions | |
AstrometryFit (std::shared_ptr< Associations > associations, std::shared_ptr< AstrometryModel > astrometryModel, double posError) | |
this is the only constructor More... | |
AstrometryFit (AstrometryFit const &)=delete | |
No copy or move: there is only ever one fitter of a given type. More... | |
AstrometryFit (AstrometryFit &&)=delete | |
AstrometryFit & | operator= (AstrometryFit const &)=delete |
AstrometryFit & | operator= (AstrometryFit &&)=delete |
void | assignIndices (std::string const &whatToFit) override |
Set parameters to fit and assign indices in the big matrix. More... | |
void | freezeErrorScales () |
The transformations used to propagate errors are freezed to the current state. More... | |
void | offsetParams (Eigen::VectorXd const &delta) override |
Offset the parameters by the requested quantities. More... | |
void | saveResultTuples (std::string const &tupleName) const override |
Save the full chi2 term per star that was used in the minimization, for debugging. More... | |
void | makeMeasResTuple (std::string const &tupleName) const |
Produces a tuple containing residuals of measurement terms. More... | |
void | makeRefResTuple (std::string const &tupleName) const |
Produces a tuple containing residuals of reference terms. More... | |
void | checkStuff () |
DEBUGGING routine. More... | |
![]() | |
FitterBase (std::shared_ptr< Associations > associations) | |
FitterBase (FitterBase const &)=delete | |
No copy or move: there is only ever one fitter of a given type. More... | |
FitterBase (FitterBase &&)=delete | |
FitterBase & | operator= (FitterBase const &)=delete |
FitterBase & | operator= (FitterBase &&)=delete |
MinimizeResult | minimize (std::string const &whatToFit, double nSigmaCut=0) |
Does a 1 step minimization, assuming a linear model. More... | |
Chi2Statistic | computeChi2 () const |
Returns the chi2 for the current state. More... | |
void | leastSquareDerivatives (TripletList &tripletList, Eigen::VectorXd &grad) const |
Evaluates the chI^2 derivatives (Jacobian and gradient) for the current whatToFit setting. More... | |
Additional Inherited Members | |
![]() | |
unsigned | findOutliers (double nSigmaCut, MeasuredStarList &msOutliers, FittedStarList &fsOutliers) const |
Find Measurements and references contributing more than a cut, computed as
\[ <chi2> + nSigmaCut + rms(chi2). \] The outliers are NOT removed, and no refit is done. More... | |
void | outliersContributions (MeasuredStarList &msOutliers, FittedStarList &fsOutliers, TripletList &tripletList, Eigen::VectorXd &grad) |
Contributions to derivatives from (presumably) outlier terms. More... | |
void | removeMeasOutliers (MeasuredStarList &outliers) |
Remove measuredStar outliers from the fit. No Refit done. More... | |
void | removeRefOutliers (FittedStarList &outliers) |
Remove refStar outliers from the fit. No Refit done. More... | |
![]() | |
std::shared_ptr< Associations > | _associations |
std::string | _whatToFit |
int | _lastNTrip |
unsigned int | _nParTot |
unsigned | _nMeasuredStars |
Class that handles the astrometric least squares problem.
This is the class that actually computes the quantities required to carry out a LS astrometric fit wrt distortion mappings and coordinates of common objects. Namely it computes the Jacobian and gradient of the chi2 (w.r.t. parameters), and the Chi2 itself. It interfaces with the actual modelling of distortions via a mimimum virtual interface AstrometryModel, and the actual mappings via an other virtual interface : Mapping.
In short AstrometryFit aims at computing derivatives of least quares. The terms of the chi2 are of two kinds:
kind 1 -> (T(X_M) - p(F))^T W (T(X_M) - p(F))
with X_M is a measured (2d) position in CCD coordinates, F refers to the position of the object in some space, defined in practise by p. There is one such term per measurement. The default setup would be that p is the projection from sky to some tangent plane and hence T maps the CCD coordinates onto this TP. p is obtained via the DistorsionModel and can be different for all CcdImage's. Depending on what is beeing fitted, one could imagine cases where the projector p is the same for all CcdImages.
Kind 2 -> (p'(F)-p'(R))^T W_R (p'(F)-p'(R)) R refers to some externally-provided reference object position, and p' to some projector from sky to some plane. The reference objects define the overall coordinate frame, which is required when all T and all F are fitted simultaneously. There is one such term per external reference object. There can be more F (fitted) objects than R (reference) objects.
In the same framework, one can fit relative transforms between images by setting p = Identity for all input CcdImages and not fitting T for one of the CcdImage's. One does not need reference object and would then naturally not have any Kind 2 terms.
Definition at line 55 of file AstrometryFit.h.
lsst::jointcal::AstrometryFit::AstrometryFit | ( | std::shared_ptr< Associations > | associations, |
std::shared_ptr< AstrometryModel > | astrometryModel, | ||
double | posError | ||
) |
this is the only constructor
Definition at line 28 of file AstrometryFit.cc.
|
delete |
No copy or move: there is only ever one fitter of a given type.
|
delete |
|
overridevirtual |
Set parameters to fit and assign indices in the big matrix.
[in] | whatToFit | Valid strings: zero or more of "Distortions", "Positions", "Refrac", "PM" which define which parameter set is going to be variable when computing derivatives (leastSquareDerivatives) and minimizing (minimize()). whatToFit="Positions Distortions" will minimize w.r.t mappings and objects positions, and not w.r.t proper motions and refraction modeling. However if proper motions and/or refraction parameters have already been set, then they are accounted for when computing residuals. The string is forwarded to the AstrometryModel, and it can then be used to turn subsets of distortion parameter on or off, if the AstrometryModel implements such a thing. |
Implements lsst::jointcal::FitterBase.
Definition at line 431 of file AstrometryFit.cc.
void lsst::jointcal::AstrometryFit::checkStuff | ( | ) |
DEBUGGING routine.
Definition at line 521 of file AstrometryFit.cc.
|
inline |
The transformations used to propagate errors are freezed to the current state.
The routine can be called when the mappings are roughly in place. After the call, the transformations used to propage errors are no longer affected when updating the mappings. This allows to have an exactly linear fit, which can be useful.
Definition at line 94 of file AstrometryFit.h.
void lsst::jointcal::AstrometryFit::makeMeasResTuple | ( | std::string const & | tupleName | ) | const |
Produces a tuple containing residuals of measurement terms.
Definition at line 567 of file AstrometryFit.cc.
void lsst::jointcal::AstrometryFit::makeRefResTuple | ( | std::string const & | tupleName | ) | const |
Produces a tuple containing residuals of reference terms.
Definition at line 625 of file AstrometryFit.cc.
|
overridevirtual |
Offset the parameters by the requested quantities.
The used parameter layout is the one from the last call to assignIndices or minimize(). There is no easy way to check that the current setting of whatToFit and the provided Delta vector are compatible: we can only test the size.
[in] | delta | vector of offsets to apply |
Implements lsst::jointcal::FitterBase.
Definition at line 469 of file AstrometryFit.cc.
|
delete |
|
delete |
|
overridevirtual |
Save the full chi2 term per star that was used in the minimization, for debugging.
Saves results to text files "tupleName-meas" and "tupleName-ref" for the MeasuredStar and RefStar tuples, respectively. This method is mostly useful for debugging: we will probably want to create a better persistence system for jointcal's internal representations in the future.
Implements lsst::jointcal::FitterBase.
Definition at line 553 of file AstrometryFit.cc.