lsst.meas.astrom.fitSipDistortion.FitSipDistortionTask Class Reference

Inheritance diagram for lsst.meas.astrom.fitSipDistortion.FitSipDistortionTask:

Public Member Functions
def	__init__ (self, **kwargs)
def	fitWcs (self, matches, initWcs, bbox=None, refCat=None, sourceCat=None, exposure=None)
def	display (self, revFitter, exposure=None, bbox=None, frame=0, pause=True)
def	makeInitialWcs (self, matches, wcs)

Public Attributes
	outlierRejectionCtrl

Static Public Attributes
	ConfigClass = FitSipDistortionConfig

Detailed Description

Fit a TAN-SIP WCS given a list of reference object/source matches.

Definition at line 101 of file fitSipDistortion.py.

Constructor & Destructor Documentation

__init__()

def lsst.meas.astrom.fitSipDistortion.FitSipDistortionTask.__init__	(		self,
		**	kwargs
	)

Definition at line 107 of file fitSipDistortion.py.

    def __init__(self, **kwargs):
        lsst.pipe.base.Task.__init__(self, **kwargs)
        self.outlierRejectionCtrl = OutlierRejectionControl()
        self.outlierRejectionCtrl.nClipMin = self.config.nClipMin
        self.outlierRejectionCtrl.nClipMax = self.config.nClipMax
        self.outlierRejectionCtrl.nSigma = self.config.rejSigma
 

Member Function Documentation

display()

def lsst.meas.astrom.fitSipDistortion.FitSipDistortionTask.display	(		self,
			revFitter,
			exposure = None,
			bbox = None,
			frame = 0,
			pause = True
	)

Display positions and outlier status overlaid on an image.

This method is called by fitWcs when display debugging is enabled.  It
always drops into pdb before returning to allow interactive inspection,
and hence it should never be called in non-interactive contexts.

Parameters
----------
revFitter : :cpp:class:`lsst::meas::astrom::ScaledPolynomialTransformFitter`
    Fitter object initialized with `fromMatches` for fitting a "reverse"
    distortion: the mapping from intermediate world coordinates to
    pixels.
exposure : :cpp:class:`lsst::afw::image::Exposure`
    An Exposure or other displayable image on which matches can be
    overplotted.
bbox : :cpp:class:`lsst::afw::geom::Box2I`
    Bounding box of the region on which matches should be plotted.

Definition at line 263 of file fitSipDistortion.py.

    def display(self, revFitter, exposure=None, bbox=None, frame=0, pause=True):
        """Display positions and outlier status overlaid on an image.
 
        This method is called by fitWcs when display debugging is enabled.  It
        always drops into pdb before returning to allow interactive inspection,
        and hence it should never be called in non-interactive contexts.
 
        Parameters
        ----------
        revFitter : :cpp:class:`lsst::meas::astrom::ScaledPolynomialTransformFitter`
            Fitter object initialized with `fromMatches` for fitting a "reverse"
            distortion: the mapping from intermediate world coordinates to
            pixels.
        exposure : :cpp:class:`lsst::afw::image::Exposure`
            An Exposure or other displayable image on which matches can be
            overplotted.
        bbox : :cpp:class:`lsst::afw::geom::Box2I`
            Bounding box of the region on which matches should be plotted.
        """
        data = revFitter.getData()
        disp = lsst.afw.display.getDisplay(frame=frame)
        if exposure is not None:
            disp.mtv(exposure)
        elif bbox is not None:
            disp.mtv(exposure=lsst.afw.image.ExposureF(bbox))
        else:
            raise TypeError("At least one of 'exposure' and 'bbox' must be provided.")
        data = revFitter.getData()
        srcKey = lsst.afw.table.Point2DKey(data.schema["src"])
        srcErrKey = lsst.afw.table.CovarianceMatrix2fKey(data.schema["src"], ["x", "y"])
        refKey = lsst.afw.table.Point2DKey(data.schema["initial"])
        modelKey = lsst.afw.table.Point2DKey(data.schema["model"])
        rejectedKey = data.schema.find("rejected").key
        with disp.Buffering():
            for record in data:
                colors = ((lsst.afw.display.RED, lsst.afw.display.GREEN)
                          if not record.get(rejectedKey) else
                          (lsst.afw.display.MAGENTA, lsst.afw.display.CYAN))
                rx, ry = record.get(refKey)
                disp.dot("x", rx, ry, size=10, ctype=colors[0])
                mx, my = record.get(modelKey)
                disp.dot("o", mx, my, size=10, ctype=colors[0])
                disp.line([(rx, ry), (mx, my)], ctype=colors[0])
                sx, sy = record.get(srcKey)
                sErr = record.get(srcErrKey)
                sEllipse = lsst.afw.geom.Quadrupole(sErr[0, 0], sErr[1, 1], sErr[0, 1])
                disp.dot(sEllipse, sx, sy, ctype=colors[1])
        if pause or pause is None:  # default is to pause
            print("Dropping into debugger to allow inspection of display. Type 'continue' when done.")
            import pdb
            pdb.set_trace()
            return frame
        else:
            return frame + 1    # increment and return the frame for the next iteration.
 

fitWcs()

def lsst.meas.astrom.fitSipDistortion.FitSipDistortionTask.fitWcs	(		self,
			matches,
			initWcs,
			bbox = None,
			refCat = None,
			sourceCat = None,
			exposure = None
	)

Fit a TAN-SIP WCS from a list of reference object/source matches.

Parameters
----------
matches : `list` of `lsst.afw.table.ReferenceMatch`
    A sequence of reference object/source matches.
    The following fields are read:
    - match.first (reference object) coord
    - match.second (source) centroid

    The following fields are written:
    - match.first (reference object) centroid
    - match.second (source) centroid
    - match.distance (on sky separation, in radians)

initWcs : `lsst.afw.geom.SkyWcs`
    An initial WCS whose CD matrix is used as the final CD matrix.
bbox : `lsst.geom.Box2I`
    The region over which the WCS will be valid (PARENT pixel coordinates);
    if `None` or an empty box then computed from matches
refCat : `lsst.afw.table.SimpleCatalog`
    Reference object catalog, or `None`.
    If provided then all centroids are updated with the new WCS,
    otherwise only the centroids for ref objects in matches are updated.
    Required fields are "centroid_x", "centroid_y", "coord_ra", and "coord_dec".
sourceCat : `lsst.afw.table.SourceCatalog`
    Source catalog, or `None`.
    If provided then coords are updated with the new WCS;
    otherwise only the coords for sources in matches are updated.
    Required input fields are "slot_Centroid_x", "slot_Centroid_y",
    "slot_Centroid_xErr", "slot_Centroid_yErr", and optionally
    "slot_Centroid_x_y_Cov".  The "coord_ra" and "coord_dec" fields
    will be updated but are not used as input.
exposure : `lsst.afw.image.Exposure`
    An Exposure or other displayable image on which matches can be
    overplotted.  Ignored (and may be `None`) if display-based debugging
    is not enabled via lsstDebug.

Returns
-------
An lsst.pipe.base.Struct with the following fields:
    - wcs : `lsst.afw.geom.SkyWcs`
        The best-fit WCS.
    - scatterOnSky : `lsst.geom.Angle`
        The median on-sky separation between reference objects and
        sources in "matches", as an `lsst.geom.Angle`

Definition at line 115 of file fitSipDistortion.py.

    def fitWcs(self, matches, initWcs, bbox=None, refCat=None, sourceCat=None, exposure=None):
        """Fit a TAN-SIP WCS from a list of reference object/source matches.
 
        Parameters
        ----------
        matches : `list` of `lsst.afw.table.ReferenceMatch`
            A sequence of reference object/source matches.
            The following fields are read:
            - match.first (reference object) coord
            - match.second (source) centroid
 
            The following fields are written:
            - match.first (reference object) centroid
            - match.second (source) centroid
            - match.distance (on sky separation, in radians)
 
        initWcs : `lsst.afw.geom.SkyWcs`
            An initial WCS whose CD matrix is used as the final CD matrix.
        bbox : `lsst.geom.Box2I`
            The region over which the WCS will be valid (PARENT pixel coordinates);
            if `None` or an empty box then computed from matches
        refCat : `lsst.afw.table.SimpleCatalog`
            Reference object catalog, or `None`.
            If provided then all centroids are updated with the new WCS,
            otherwise only the centroids for ref objects in matches are updated.
            Required fields are "centroid_x", "centroid_y", "coord_ra", and "coord_dec".
        sourceCat : `lsst.afw.table.SourceCatalog`
            Source catalog, or `None`.
            If provided then coords are updated with the new WCS;
            otherwise only the coords for sources in matches are updated.
            Required input fields are "slot_Centroid_x", "slot_Centroid_y",
            "slot_Centroid_xErr", "slot_Centroid_yErr", and optionally
            "slot_Centroid_x_y_Cov".  The "coord_ra" and "coord_dec" fields
            will be updated but are not used as input.
        exposure : `lsst.afw.image.Exposure`
            An Exposure or other displayable image on which matches can be
            overplotted.  Ignored (and may be `None`) if display-based debugging
            is not enabled via lsstDebug.
 
        Returns
        -------
        An lsst.pipe.base.Struct with the following fields:
            - wcs : `lsst.afw.geom.SkyWcs`
                The best-fit WCS.
            - scatterOnSky : `lsst.geom.Angle`
                The median on-sky separation between reference objects and
                sources in "matches", as an `lsst.geom.Angle`
        """
        import lsstDebug
        display = lsstDebug.Info(__name__).display
        displayFrame = lsstDebug.Info(__name__).frame
        displayPause = lsstDebug.Info(__name__).pause
 
        if bbox is None:
            bbox = lsst.geom.Box2D()
            for match in matches:
                bbox.include(match.second.getCentroid())
            bbox = lsst.geom.Box2I(bbox)
 
        wcs = self.makeInitialWcs(matches, initWcs)
        cdMatrix = lsst.geom.LinearTransform(wcs.getCdMatrix())
 
        # Fit the "reverse" mapping from intermediate world coordinates to
        # pixels, rejecting outliers. Fitting in this direction first makes it
        # easier to handle the case where we have uncertainty on source
        # positions but not reference positions.  That's the case we have
        # right now for purely bookeeeping reasons, and it may be the case we
        # have in the future when we us Gaia as the reference catalog.
        revFitter = ScaledPolynomialTransformFitter.fromMatches(self.config.order, matches, wcs,
                                                                self.config.refUncertainty)
        revFitter.fit()
        for nIter in range(self.config.numRejIter):
            revFitter.updateModel()
            intrinsicScatter = revFitter.updateIntrinsicScatter()
            clippedSigma, nRejected = revFitter.rejectOutliers(self.outlierRejectionCtrl)
            self.log.debug(
                "Iteration {0}: intrinsic scatter is {1:4.3f} pixels, "
                "rejected {2} outliers at {3:3.2f} sigma.".format(
                    nIter+1, intrinsicScatter, nRejected, clippedSigma
                )
            )
            if display:
                displayFrame = self.display(revFitter, exposure=exposure, bbox=bbox,
                                            frame=displayFrame, displayPause=displayPause)
            revFitter.fit()
        revScaledPoly = revFitter.getTransform()
        # Convert the generic ScaledPolynomialTransform result to SIP form
        # with given CRPIX and CD (this is an exact conversion, up to
        # floating-point round-off error)
        sipReverse = SipReverseTransform.convert(revScaledPoly, wcs.getPixelOrigin(), cdMatrix)
 
        # Fit the forward mapping to a grid of points created from the reverse
        # transform.  Because that grid needs to be defined in intermediate
        # world coordinates, and we don't have a good way to get from pixels to
        # intermediate world coordinates yet (that's what we're fitting), we'll
        # first grow the box to make it conservatively large...
        gridBBoxPix = lsst.geom.Box2D(bbox)
        gridBBoxPix.grow(self.config.gridBorder)
        # ...and then we'll transform using just the CRPIX offset and CD matrix
        # linear transform, which is the TAN-only (no SIP distortion, and
        # hence approximate) mapping from pixels to intermediate world
        # coordinates.
        gridBBoxIwc = lsst.geom.Box2D()
        for point in gridBBoxPix.getCorners():
            point -= lsst.geom.Extent2D(wcs.getPixelOrigin())
            gridBBoxIwc.include(cdMatrix(point))
        fwdFitter = ScaledPolynomialTransformFitter.fromGrid(self.config.order, gridBBoxIwc,
                                                             self.config.nGridX, self.config.nGridY,
                                                             revScaledPoly)
        fwdFitter.fit()
        # Convert to SIP forward form.
        fwdScaledPoly = fwdFitter.getTransform()
        sipForward = SipForwardTransform.convert(fwdScaledPoly, wcs.getPixelOrigin(), cdMatrix)
 
        # Make a new WCS from the SIP transform objects and the CRVAL in the
        # initial WCS.
        wcs = makeWcs(sipForward, sipReverse, wcs.getSkyOrigin())
 
        if refCat is not None:
            self.log.debug("Updating centroids in refCat")
            lsst.afw.table.updateRefCentroids(wcs, refList=refCat)
        else:
            self.log.warn("Updating reference object centroids in match list; refCat is None")
            lsst.afw.table.updateRefCentroids(wcs, refList=[match.first for match in matches])
 
        if sourceCat is not None:
            self.log.debug("Updating coords in sourceCat")
            lsst.afw.table.updateSourceCoords(wcs, sourceList=sourceCat)
        else:
            self.log.warn("Updating source coords in match list; sourceCat is None")
            lsst.afw.table.updateSourceCoords(wcs, sourceList=[match.second for match in matches])
 
        self.log.debug("Updating distance in match list")
        setMatchDistance(matches)
 
        stats = makeMatchStatisticsInRadians(wcs, matches, lsst.afw.math.MEDIAN)
        scatterOnSky = stats.getValue()*lsst.geom.radians
 
        if scatterOnSky.asArcseconds() > self.config.maxScatterArcsec:
            raise lsst.pipe.base.TaskError(
                "Fit failed: median scatter on sky = %0.3f arcsec > %0.3f config.maxScatterArcsec" %
                (scatterOnSky.asArcseconds(), self.config.maxScatterArcsec))
 
        return lsst.pipe.base.Struct(
            wcs=wcs,
            scatterOnSky=scatterOnSky,
        )
 

makeInitialWcs()

def lsst.meas.astrom.fitSipDistortion.FitSipDistortionTask.makeInitialWcs	(		self,
			matches,
			wcs
	)

Generate a guess Wcs from the astrometric matches

We create a Wcs anchored at the center of the matches, with the scale
of the input Wcs.  This is necessary because the Wcs may have a very
approximation position (as is common with telescoped-generated Wcs).
We're using the best of each: positions from the matches, and scale
from the input Wcs.

Parameters
----------
matches : list of :cpp:class:`lsst::afw::table::ReferenceMatch`
    A sequence of reference object/source matches.
    The following fields are read:

    - match.first (reference object) coord
    - match.second (source) centroid

wcs : :cpp:class:`lsst::afw::geom::SkyWcs`
    An initial WCS whose CD matrix is used as the CD matrix of the
    result.

Returns
-------
newWcs : `lsst.afw.geom.SkyWcs`
    A new WCS guess.

Definition at line 318 of file fitSipDistortion.py.

    def makeInitialWcs(self, matches, wcs):
        """Generate a guess Wcs from the astrometric matches
 
        We create a Wcs anchored at the center of the matches, with the scale
        of the input Wcs.  This is necessary because the Wcs may have a very
        approximation position (as is common with telescoped-generated Wcs).
        We're using the best of each: positions from the matches, and scale
        from the input Wcs.
 
        Parameters
        ----------
        matches : list of :cpp:class:`lsst::afw::table::ReferenceMatch`
            A sequence of reference object/source matches.
            The following fields are read:
 
            - match.first (reference object) coord
            - match.second (source) centroid
 
        wcs : :cpp:class:`lsst::afw::geom::SkyWcs`
            An initial WCS whose CD matrix is used as the CD matrix of the
            result.
 
        Returns
        -------
        newWcs : `lsst.afw.geom.SkyWcs`
            A new WCS guess.
        """
        crpix = lsst.geom.Extent2D(0, 0)
        crval = lsst.sphgeom.Vector3d(0, 0, 0)
        for mm in matches:
            crpix += lsst.geom.Extent2D(mm.second.getCentroid())
            crval += mm.first.getCoord().getVector()
        crpix /= len(matches)
        crval /= len(matches)
        cd = wcs.getCdMatrix()
        newWcs = lsst.afw.geom.makeSkyWcs(crpix=lsst.geom.Point2D(crpix),
                                          crval=lsst.geom.SpherePoint(crval),
                                          cdMatrix=cd)
        return newWcs

Member Data Documentation

ConfigClass

lsst.meas.astrom.fitSipDistortion.FitSipDistortionTask.ConfigClass = FitSipDistortionConfig

static

Definition at line 104 of file fitSipDistortion.py.

outlierRejectionCtrl

lsst.meas.astrom.fitSipDistortion.FitSipDistortionTask.outlierRejectionCtrl

Definition at line 109 of file fitSipDistortion.py.

---

The documentation for this class was generated from the following file:

• /j/snowflake/release/lsstsw/stack/lsst-scipipe-0.7.0/Linux64/meas_astrom/22.0.1-4-g8623105+b8044ec9de/python/lsst/meas/astrom/fitSipDistortion.py