Inheritance diagram for lsst.ip.diffim.imageDecorrelation.DecorrelateALKernelTask:

Public Member Functions
def	__init__ (self, args, *kwargs)

def	computeVarianceMean (self, exposure)

def	run (self, scienceExposure, templateExposure, subtractedExposure, psfMatchingKernel, preConvKernel=None, xcen=None, ycen=None, svar=None, tvar=None, templateMatched=True, preConvMode=False, **kwargs)

def	computeCommonShape (self, *shapes)

def	computeDiffimCorrection (self, kappa, svar, tvar)

def	computeScoreCorrection (self, kappa, svar, tvar, preConvArr)

def	calculateVariancePlane (self, vplane1, vplane2, varMean1, varMean2, c1ft, c2ft)

def	computeCorrectedDiffimPsf (self, corrft, psfOld)

def	computeCorrectedImage (self, corrft, imgOld)

Static Public Member Functions
def	padCenterOriginArray (A, tuple newShape, useInverse=False)

def	estimateVariancePlane (vplane1, vplane2, c1ft, c2ft)

Public Attributes
	statsControl

	freqSpaceShape

Static Public Attributes
	ConfigClass = DecorrelateALKernelConfig

Detailed Description

Decorrelate the effect of convolution by Alard-Lupton matching kernel in image difference

Notes
-----

Pipe-task that removes the neighboring-pixel covariance in an
image difference that are added when the template image is
convolved with the Alard-Lupton PSF matching kernel.

The image differencing pipeline task @link
ip.diffim.psfMatch.PsfMatchTask PSFMatchTask@endlink and @link
ip.diffim.psfMatch.PsfMatchConfigAL PSFMatchConfigAL@endlink uses
the Alard and Lupton (1998) method for matching the PSFs of the
template and science exposures prior to subtraction. The
Alard-Lupton method identifies a matching kernel, which is then
(typically) convolved with the template image to perform PSF
matching. This convolution has the effect of adding covariance
between neighboring pixels in the template image, which is then
added to the image difference by subtraction.

The pixel covariance may be corrected by whitening the noise of
the image difference. This task performs such a decorrelation by
computing a decorrelation kernel (based upon the A&L matching
kernel and variances in the template and science images) and
convolving the image difference with it. This process is described
in detail in [DMTN-021](http://dmtn-021.lsst.io).

This task has no standalone example, however it is applied as a
subtask of pipe.tasks.imageDifference.ImageDifferenceTask.

Definition at line 59 of file imageDecorrelation.py.

Constructor & Destructor Documentation

◆ init()

def lsst.ip.diffim.imageDecorrelation.DecorrelateALKernelTask.__init__	(		self,
		*	args,
		**	kwargs
	)

Create the image decorrelation Task

Parameters
----------
args :
    arguments to be passed to ``lsst.pipe.base.task.Task.__init__``
kwargs :
    keyword arguments to be passed to ``lsst.pipe.base.task.Task.__init__``

Reimplemented in lsst.ip.diffim.imageDecorrelation.DecorrelateALKernelMapper.

Definition at line 93 of file imageDecorrelation.py.

     def __init__(self, *args, **kwargs):
         """Create the image decorrelation Task
  
         Parameters
         ----------
         args :
             arguments to be passed to ``lsst.pipe.base.task.Task.__init__``
         kwargs :
             keyword arguments to be passed to ``lsst.pipe.base.task.Task.__init__``
         """
         pipeBase.Task.__init__(self, *args, **kwargs)
  
         self.statsControl = afwMath.StatisticsControl()
         self.statsControl.setNumSigmaClip(3.)
         self.statsControl.setNumIter(3)
         self.statsControl.setAndMask(afwImage.Mask.getPlaneBitMask(self.config.ignoreMaskPlanes))
  

Member Function Documentation

◆ calculateVariancePlane()

def lsst.ip.diffim.imageDecorrelation.DecorrelateALKernelTask.calculateVariancePlane	(	self,
		vplane1,
		vplane2,
		varMean1,
		varMean2,
		c1ft,
		c2ft
	)

Full propagation of the variance planes of the original exposures.

The original variance planes of independent pixels are convolved with the
image space square of the overall kernels.

Parameters
----------
vplane1, vplane2 : `numpy.ndarray` of `float`
    Variance planes of the original (before pre-convolution or matching)
    exposures.
varMean1, varMean2 : `float`
    Replacement average values for non-finite ``vplane1`` and ``vplane2`` values respectively.

c1ft, c2ft : `numpy.ndarray` of `complex`
    The overall convolution that includes the matching and the
    afterburner in frequency space. The result of either
    ``computeScoreCorrection`` or ``computeDiffimCorrection``.

Returns
-------
vplaneD : `numpy.ndarray` of `float`
  The variance plane of the difference/score images.

Notes
------
See DMTN-179 Section 5 about the variance plane calculations.

Infs and NaNs are allowed and kept in the returned array.

Definition at line 537 of file imageDecorrelation.py.

     def calculateVariancePlane(self, vplane1, vplane2, varMean1, varMean2, c1ft, c2ft):
         """Full propagation of the variance planes of the original exposures.
  
         The original variance planes of independent pixels are convolved with the
         image space square of the overall kernels.
  
         Parameters
         ----------
         vplane1, vplane2 : `numpy.ndarray` of `float`
             Variance planes of the original (before pre-convolution or matching)
             exposures.
         varMean1, varMean2 : `float`
             Replacement average values for non-finite ``vplane1`` and ``vplane2`` values respectively.
  
         c1ft, c2ft : `numpy.ndarray` of `complex`
             The overall convolution that includes the matching and the
             afterburner in frequency space. The result of either
             ``computeScoreCorrection`` or ``computeDiffimCorrection``.
  
         Returns
         -------
         vplaneD : `numpy.ndarray` of `float`
           The variance plane of the difference/score images.
  
         Notes
         ------
         See DMTN-179 Section 5 about the variance plane calculations.
  
         Infs and NaNs are allowed and kept in the returned array.
         """
         D = np.real(np.fft.ifft2(c1ft))
         c1SqFt = np.fft.fft2(D*D)
  
         v1shape = vplane1.shape
         filtInf = np.isinf(vplane1)
         filtNan = np.isnan(vplane1)
         # This copy could be eliminated if inf/nan handling were go into padCenterOriginArray
         vplane1 = np.copy(vplane1)
         vplane1[filtInf | filtNan] = varMean1
         D = self.padCenterOriginArray(vplane1, self.freqSpaceShape)
         v1 = np.real(np.fft.ifft2(np.fft.fft2(D) * c1SqFt))
         v1 = self.padCenterOriginArray(v1, v1shape, useInverse=True)
         v1[filtNan] = np.nan
         v1[filtInf] = np.inf
  
         D = np.real(np.fft.ifft2(c2ft))
         c2ft = np.fft.fft2(D*D)
  
         v2shape = vplane2.shape
         filtInf = np.isinf(vplane2)
         filtNan = np.isnan(vplane2)
         vplane2 = np.copy(vplane2)
         vplane2[filtInf | filtNan] = varMean2
         D = self.padCenterOriginArray(vplane2, self.freqSpaceShape)
         v2 = np.real(np.fft.ifft2(np.fft.fft2(D) * c2ft))
         v2 = self.padCenterOriginArray(v2, v2shape, useInverse=True)
         v2[filtNan] = np.nan
         v2[filtInf] = np.inf
  
         return v1 + v2
  

◆ computeCommonShape()

def lsst.ip.diffim.imageDecorrelation.DecorrelateALKernelTask.computeCommonShape	(		self,
		*	shapes
	)

Calculate the common shape for FFT operations. Set `self.freqSpaceShape`
internally.

Parameters
----------
shapes : one or more `tuple` of `int`
    Shapes of the arrays. All must have the same dimensionality.
    At least one shape must be provided.

Returns
-------
None.

Notes
-----
For each dimension, gets the smallest even number greater than or equal to
`N1+N2-1` where `N1` and `N2` are the two largest values.
In case of only one shape given, rounds up to even each dimension value.

Definition at line 330 of file imageDecorrelation.py.

     def computeCommonShape(self, *shapes):
         """Calculate the common shape for FFT operations. Set `self.freqSpaceShape`
         internally.
  
         Parameters
         ----------
         shapes : one or more `tuple` of `int`
             Shapes of the arrays. All must have the same dimensionality.
             At least one shape must be provided.
  
         Returns
         -------
         None.
  
         Notes
         -----
         For each dimension, gets the smallest even number greater than or equal to
         `N1+N2-1` where `N1` and `N2` are the two largest values.
         In case of only one shape given, rounds up to even each dimension value.
         """
         S = np.array(shapes, dtype=int)
         if len(shapes) > 2:
             S.sort(axis=0)
             S = S[-2:]
         if len(shapes) > 1:
             commonShape = np.sum(S, axis=0) - 1
         else:
             commonShape = S[0]
         commonShape[commonShape % 2 != 0] += 1
         self.freqSpaceShape = tuple(commonShape)
         self.log.info("Common frequency space shape %s", self.freqSpaceShape)
  

◆ computeCorrectedDiffimPsf()

def lsst.ip.diffim.imageDecorrelation.DecorrelateALKernelTask.computeCorrectedDiffimPsf	(	self,
		corrft,
		psfOld
	)

Compute the (decorrelated) difference image's new PSF.

Parameters
----------
corrft : `numpy.ndarray`
    The frequency space representation of the correction calculated by
    `computeCorrection`. Shape must be `self.freqSpaceShape`.
psfOld : `numpy.ndarray`
    The psf of the difference image to be corrected.

Returns
-------
psfNew : `numpy.ndarray`
    The corrected psf, same shape as `psfOld`, sum normed to 1.

Notes
-----
There is no algorithmic guarantee that the corrected psf can
meaningfully fit to the same size as the original one.

Definition at line 598 of file imageDecorrelation.py.

     def computeCorrectedDiffimPsf(self, corrft, psfOld):
         """Compute the (decorrelated) difference image's new PSF.
  
         Parameters
         ----------
         corrft : `numpy.ndarray`
             The frequency space representation of the correction calculated by
             `computeCorrection`. Shape must be `self.freqSpaceShape`.
         psfOld : `numpy.ndarray`
             The psf of the difference image to be corrected.
  
         Returns
         -------
         psfNew : `numpy.ndarray`
             The corrected psf, same shape as `psfOld`, sum normed to 1.
  
         Notes
         -----
         There is no algorithmic guarantee that the corrected psf can
         meaningfully fit to the same size as the original one.
         """
         psfShape = psfOld.shape
         psfNew = self.padCenterOriginArray(psfOld, self.freqSpaceShape)
         psfNew = np.fft.fft2(psfNew)
         psfNew *= corrft
         psfNew = np.fft.ifft2(psfNew)
         psfNew = psfNew.real
         psfNew = self.padCenterOriginArray(psfNew, psfShape, useInverse=True)
         psfNew = psfNew/psfNew.sum()
         return psfNew
  

◆ computeCorrectedImage()

def lsst.ip.diffim.imageDecorrelation.DecorrelateALKernelTask.computeCorrectedImage	(	self,
		corrft,
		imgOld
	)

Compute the decorrelated difference image.

Parameters
----------
corrft : `numpy.ndarray`
    The frequency space representation of the correction calculated by
    `computeCorrection`. Shape must be `self.freqSpaceShape`.
imgOld : `numpy.ndarray`
    The difference image to be corrected.

Returns
-------
imgNew : `numpy.ndarray`
    The corrected image, same size as the input.

Definition at line 629 of file imageDecorrelation.py.

     def computeCorrectedImage(self, corrft, imgOld):
         """Compute the decorrelated difference image.
  
         Parameters
         ----------
         corrft : `numpy.ndarray`
             The frequency space representation of the correction calculated by
             `computeCorrection`. Shape must be `self.freqSpaceShape`.
         imgOld : `numpy.ndarray`
             The difference image to be corrected.
  
         Returns
         -------
         imgNew : `numpy.ndarray`
             The corrected image, same size as the input.
         """
         expShape = imgOld.shape
         imgNew = np.copy(imgOld)
         filtInf = np.isinf(imgNew)
         filtNan = np.isnan(imgNew)
         imgNew[filtInf] = np.nan
         imgNew[filtInf | filtNan] = np.nanmean(imgNew)
         imgNew = self.padCenterOriginArray(imgNew, self.freqSpaceShape)
         imgNew = np.fft.fft2(imgNew)
         imgNew *= corrft
         imgNew = np.fft.ifft2(imgNew)
         imgNew = imgNew.real
         imgNew = self.padCenterOriginArray(imgNew, expShape, useInverse=True)
         imgNew[filtNan] = np.nan
         imgNew[filtInf] = np.inf
         return imgNew
  
  

◆ computeDiffimCorrection()

def lsst.ip.diffim.imageDecorrelation.DecorrelateALKernelTask.computeDiffimCorrection	(	self,
		kappa,
		svar,
		tvar
	)

Compute the Lupton decorrelation post-convolution kernel for decorrelating an
image difference, based on the PSF-matching kernel.

Parameters
----------
kappa : `numpy.ndarray` of `float`
    A matching kernel 2-d numpy.array derived from Alard & Lupton PSF matching.
svar : `float` > 0.
    Average variance of science image used for PSF matching.
tvar : `float` > 0.
    Average variance of the template (matched) image used for PSF matching.

Returns
-------
corrft : `numpy.ndarray` of `float`
    The frequency space representation of the correction. The array is real (dtype float).
    Shape is `self.freqSpaceShape`.

cnft, crft : `numpy.ndarray` of `complex`
    The overall convolution (pre-conv, PSF matching, noise correction) kernel
    for the science and template images, respectively for the variance plane
    calculations. These are intermediate results in frequency space.

Notes
-----
The maximum correction factor converges to `sqrt(tvar/svar)` towards high frequencies.
This should be a plausible value.

Definition at line 410 of file imageDecorrelation.py.

     def computeDiffimCorrection(self, kappa, svar, tvar):
         """Compute the Lupton decorrelation post-convolution kernel for decorrelating an
         image difference, based on the PSF-matching kernel.
  
         Parameters
         ----------
         kappa : `numpy.ndarray` of `float`
             A matching kernel 2-d numpy.array derived from Alard & Lupton PSF matching.
         svar : `float` > 0.
             Average variance of science image used for PSF matching.
         tvar : `float` > 0.
             Average variance of the template (matched) image used for PSF matching.
  
         Returns
         -------
         corrft : `numpy.ndarray` of `float`
             The frequency space representation of the correction. The array is real (dtype float).
             Shape is `self.freqSpaceShape`.
  
         cnft, crft : `numpy.ndarray` of `complex`
             The overall convolution (pre-conv, PSF matching, noise correction) kernel
             for the science and template images, respectively for the variance plane
             calculations. These are intermediate results in frequency space.
  
         Notes
         -----
         The maximum correction factor converges to `sqrt(tvar/svar)` towards high frequencies.
         This should be a plausible value.
         """
         kSum = np.sum(kappa)  # We scale the decorrelation to preserve fluxes
         kappa = self.padCenterOriginArray(kappa, self.freqSpaceShape)
         kft = np.fft.fft2(kappa)
         kftAbsSq = np.real(np.conj(kft) * kft)
  
         denom = svar + tvar * kftAbsSq
         corrft = np.sqrt((svar + tvar * kSum*kSum) / denom)
         cnft = corrft
         crft = kft*corrft
         return pipeBase.Struct(corrft=corrft, cnft=cnft, crft=crft)
  

◆ computeScoreCorrection()

def lsst.ip.diffim.imageDecorrelation.DecorrelateALKernelTask.computeScoreCorrection	(	self,
		kappa,
		svar,
		tvar,
		preConvArr
	)

Compute the correction kernel for a score image.

Parameters
----------
kappa : `numpy.ndarray`
    A matching kernel 2-d numpy.array derived from Alard & Lupton PSF matching.
svar : `float`
    Average variance of science image used for PSF matching (before pre-convolution).
tvar : `float`
    Average variance of the template (matched) image used for PSF matching.
preConvArr : `numpy.ndarray`
    The pre-convolution kernel of the science image. It should be the PSF
    of the science image or an approximation of it. It must be normed to sum 1.

Returns
-------
corrft : `numpy.ndarray` of `float`
    The frequency space representation of the correction. The array is real (dtype float).
    Shape is `self.freqSpaceShape`.
cnft, crft : `numpy.ndarray` of `complex`
    The overall convolution (pre-conv, PSF matching, noise correction) kernel
    for the science and template images, respectively for the variance plane
    calculations. These are intermediate results in frequency space.

Notes
-----
To be precise, the science image should be _correlated_ by ``preConvArray`` but this
does not matter for this calculation.

``cnft``, ``crft`` contain the scaling factor as well.

Definition at line 450 of file imageDecorrelation.py.

     def computeScoreCorrection(self, kappa, svar, tvar, preConvArr):
         """Compute the correction kernel for a score image.
  
         Parameters
         ----------
         kappa : `numpy.ndarray`
             A matching kernel 2-d numpy.array derived from Alard & Lupton PSF matching.
         svar : `float`
             Average variance of science image used for PSF matching (before pre-convolution).
         tvar : `float`
             Average variance of the template (matched) image used for PSF matching.
         preConvArr : `numpy.ndarray`
             The pre-convolution kernel of the science image. It should be the PSF
             of the science image or an approximation of it. It must be normed to sum 1.
  
         Returns
         -------
         corrft : `numpy.ndarray` of `float`
             The frequency space representation of the correction. The array is real (dtype float).
             Shape is `self.freqSpaceShape`.
         cnft, crft : `numpy.ndarray` of `complex`
             The overall convolution (pre-conv, PSF matching, noise correction) kernel
             for the science and template images, respectively for the variance plane
             calculations. These are intermediate results in frequency space.
  
         Notes
         -----
         To be precise, the science image should be _correlated_ by ``preConvArray`` but this
         does not matter for this calculation.
  
         ``cnft``, ``crft`` contain the scaling factor as well.
  
         """
         kSum = np.sum(kappa)
         kappa = self.padCenterOriginArray(kappa, self.freqSpaceShape)
         kft = np.fft.fft2(kappa)
         preConvArr = self.padCenterOriginArray(preConvArr, self.freqSpaceShape)
         preFt = np.fft.fft2(preConvArr)
         preFtAbsSq = np.real(np.conj(preFt) * preFt)
         kftAbsSq = np.real(np.conj(kft) * kft)
         # Avoid zero division, though we don't normally approach `tiny`.
         # We have numerical noise instead.
         tiny = np.finfo(preFtAbsSq.dtype).tiny * 1000.
         flt = preFtAbsSq < tiny
         # If we pre-convolve to avoid deconvolution in AL, then kftAbsSq / preFtAbsSq
         # theoretically expected to diverge to +inf. But we don't care about the convergence
         # properties here, S goes to 0 at these frequencies anyway.
         preFtAbsSq[flt] = tiny
         denom = svar + tvar * kftAbsSq / preFtAbsSq
         corrft = (svar + tvar * kSum*kSum) / denom
         cnft = np.conj(preFt)*corrft
         crft = kft*corrft
         return pipeBase.Struct(corrft=corrft, cnft=cnft, crft=crft)
  

◆ computeVarianceMean()

def lsst.ip.diffim.imageDecorrelation.DecorrelateALKernelTask.computeVarianceMean	(	self,
		exposure
	)

Definition at line 110 of file imageDecorrelation.py.

     def computeVarianceMean(self, exposure):
         statObj = afwMath.makeStatistics(exposure.getMaskedImage().getVariance(),
                                          exposure.getMaskedImage().getMask(),
                                          afwMath.MEANCLIP, self.statsControl)
         var = statObj.getValue(afwMath.MEANCLIP)
         return var
  

◆ estimateVariancePlane()

def lsst.ip.diffim.imageDecorrelation.DecorrelateALKernelTask.estimateVariancePlane	(	vplane1,
		vplane2,
		c1ft,
		c2ft
	)

static

Estimate the variance planes.

The estimation assumes that around each pixel the surrounding
pixels' sigmas within the convolution kernel are the same.

Parameters
----------
vplane1, vplane2 : `numpy.ndarray` of `float`
    Variance planes of the original (before pre-convolution or matching)
    exposures.
c1ft, c2ft : `numpy.ndarray` of `complex`
    The overall convolution that includes the matching and the
    afterburner in frequency space. The result of either
    ``computeScoreCorrection`` or ``computeDiffimCorrection``.

Returns
-------
vplaneD : `numpy.ndarray` of `float`
  The estimated variance plane of the difference/score image
  as a weighted sum of the input variances.

Notes
------
See DMTN-179 Section 5 about the variance plane calculations.

Definition at line 505 of file imageDecorrelation.py.

     def estimateVariancePlane(vplane1, vplane2, c1ft, c2ft):
         """Estimate the variance planes.
  
         The estimation assumes that around each pixel the surrounding
         pixels' sigmas within the convolution kernel are the same.
  
         Parameters
         ----------
         vplane1, vplane2 : `numpy.ndarray` of `float`
             Variance planes of the original (before pre-convolution or matching)
             exposures.
         c1ft, c2ft : `numpy.ndarray` of `complex`
             The overall convolution that includes the matching and the
             afterburner in frequency space. The result of either
             ``computeScoreCorrection`` or ``computeDiffimCorrection``.
  
         Returns
         -------
         vplaneD : `numpy.ndarray` of `float`
           The estimated variance plane of the difference/score image
           as a weighted sum of the input variances.
  
         Notes
         ------
         See DMTN-179 Section 5 about the variance plane calculations.
         """
         w1 = np.sum(np.real(np.conj(c1ft)*c1ft)) / c1ft.size
         w2 = np.sum(np.real(np.conj(c2ft)*c2ft)) / c2ft.size
         # w1, w2: the frequency space sum of abs(c1)^2 is the same as in image
         # space.
         return vplane1*w1 + vplane2*w2
  

◆ padCenterOriginArray()

def lsst.ip.diffim.imageDecorrelation.DecorrelateALKernelTask.padCenterOriginArray	(		A,
		tuple	newShape,
			useInverse = `False`
	)

static

Zero pad an image where the origin is at the center and replace the
origin to the corner as required by the periodic input of FFT. Implement also
the inverse operation, crop the padding and re-center data.

Parameters
----------
A : `numpy.ndarray`
    An array to copy from.
newShape : `tuple` of `int`
    The dimensions of the resulting array. For padding, the resulting array
    must be larger than A in each dimension. For the inverse operation this
    must be the original, before padding size of the array.
useInverse : bool, optional
    Selector of forward, add padding, operation (False)
    or its inverse, crop padding, operation (True).

Returns
-------
R : `numpy.ndarray`
    The padded or unpadded array with shape of `newShape` and the same dtype as A.

Notes
-----
For odd dimensions, the splitting is rounded to
put the center pixel into the new corner origin (0,0). This is to be consistent
e.g. for a dirac delta kernel that is originally located at the center pixel.

Definition at line 363 of file imageDecorrelation.py.

     def padCenterOriginArray(A, newShape: tuple, useInverse=False):
         """Zero pad an image where the origin is at the center and replace the
         origin to the corner as required by the periodic input of FFT. Implement also
         the inverse operation, crop the padding and re-center data.
  
         Parameters
         ----------
         A : `numpy.ndarray`
             An array to copy from.
         newShape : `tuple` of `int`
             The dimensions of the resulting array. For padding, the resulting array
             must be larger than A in each dimension. For the inverse operation this
             must be the original, before padding size of the array.
         useInverse : bool, optional
             Selector of forward, add padding, operation (False)
             or its inverse, crop padding, operation (True).
  
         Returns
         -------
         R : `numpy.ndarray`
             The padded or unpadded array with shape of `newShape` and the same dtype as A.
  
         Notes
         -----
         For odd dimensions, the splitting is rounded to
         put the center pixel into the new corner origin (0,0). This is to be consistent
         e.g. for a dirac delta kernel that is originally located at the center pixel.
         """
  
         # The forward and inverse operations should round odd dimension halves at the opposite
         # sides to get the pixels back to their original positions.
         if not useInverse:
             # Forward operation: First and second halves with respect to the axes of A.
             firstHalves = [x//2 for x in A.shape]
             secondHalves = [x-y for x, y in zip(A.shape, firstHalves)]
         else:
             # Inverse operation: Opposite rounding
             secondHalves = [x//2 for x in newShape]
             firstHalves = [x-y for x, y in zip(newShape, secondHalves)]
  
         R = np.zeros_like(A, shape=newShape)
         R[-firstHalves[0]:, -firstHalves[1]:] = A[:firstHalves[0], :firstHalves[1]]
         R[:secondHalves[0], -firstHalves[1]:] = A[-secondHalves[0]:, :firstHalves[1]]
         R[:secondHalves[0], :secondHalves[1]] = A[-secondHalves[0]:, -secondHalves[1]:]
         R[-firstHalves[0]:, :secondHalves[1]] = A[:firstHalves[0], -secondHalves[1]:]
         return R
  

◆ run()

def lsst.ip.diffim.imageDecorrelation.DecorrelateALKernelTask.run	(		self,
			scienceExposure,
			templateExposure,
			subtractedExposure,
			psfMatchingKernel,
			preConvKernel = `None`,
			xcen = `None`,
			ycen = `None`,
			svar = `None`,
			tvar = `None`,
			templateMatched = `True`,
			preConvMode = `False`,
		**	kwargs
	)

Perform decorrelation of an image difference or of a score difference exposure.

Corrects the difference or score image due to the convolution of the
templateExposure with the A&L PSF matching kernel.
See [DMTN-021, Equation 1](http://dmtn-021.lsst.io/#equation-1) and
[DMTN-179](http://dmtn-179.lsst.io/) for details.

Parameters
----------
scienceExposure : `lsst.afw.image.Exposure`
The original science exposure (before pre-convolution, if ``preConvMode==True``).
templateExposure : `lsst.afw.image.Exposure`
The original template exposure warped into the science exposure dimensions.
subtractedExposure : `lsst.afw.image.Exposure`
the subtracted exposure produced by
`ip_diffim.ImagePsfMatchTask.subtractExposures()`. The `subtractedExposure` must
inherit its PSF from `exposure`, see notes below.
psfMatchingKernel : `lsst.afw.detection.Psf`
An (optionally spatially-varying) PSF matching kernel produced
by `ip_diffim.ImagePsfMatchTask.subtractExposures()`.
preConvKernel : `lsst.afw.math.Kernel`, optional
If not `None`, then the `scienceExposure` was pre-convolved with (the reflection of)
this kernel. Must be normalized to sum to 1.
Allowed only if ``templateMatched==True`` and ``preConvMode==True``.
Defaults to the PSF of the science exposure at the image center.
xcen : `float`, optional
X-pixel coordinate to use for computing constant matching kernel to use
If `None` (default), then use the center of the image.
ycen : `float`, optional
Y-pixel coordinate to use for computing constant matching kernel to use
If `None` (default), then use the center of the image.
svar : `float`, optional
Image variance for science image
If `None` (default) then compute the variance over the entire input science image.
tvar : `float`, optional
Image variance for template image
If `None` (default) then compute the variance over the entire input template image.
templateMatched : `bool`, optional
If True, the template exposure was matched (convolved) to the science exposure.
See also notes below.
preConvMode : `bool`, optional
If True, ``subtractedExposure`` is assumed to be a likelihood difference image
and will be noise corrected as a likelihood image.
**kwargs
Additional keyword arguments propagated from DecorrelateALKernelSpatialTask.

Returns
-------
result : `lsst.pipe.base.Struct`
- ``correctedExposure`` : the decorrelated diffim

Notes
-----
If ``preConvMode==True``, ``subtractedExposure`` is assumed to be a
score image and the noise correction for likelihood images
is applied. The resulting image is an optimal detection likelihood image
when the templateExposure has noise. (See DMTN-179) If ``preConvKernel`` is
not specified, the PSF of ``scienceExposure`` is assumed as pre-convolution kernel.

The ``subtractedExposure`` is NOT updated. The returned ``correctedExposure``
has an updated but spatially fixed PSF. It is calculated as the center of
image PSF corrected by the center of image matching kernel.

If ``templateMatched==True``, the templateExposure was matched (convolved)
to the ``scienceExposure`` by ``psfMatchingKernel``. Otherwise the ``scienceExposure``
was matched (convolved) by ``psfMatchingKernel``.

This task discards the variance plane of ``subtractedExposure`` and re-computes
it from the variance planes of ``scienceExposure`` and ``templateExposure``.
The image plane of ``subtractedExposure`` must be at the photometric level
set by the AL PSF matching in `ImagePsfMatchTask.subtractExposures`.
The assumptions about the photometric level are controlled by the
`templateMatched` option in this task.

Here we currently convert a spatially-varying matching kernel into a constant kernel,
just by computing it at the center of the image (tickets DM-6243, DM-6244).

We are also using a constant accross-the-image measure of sigma (sqrt(variance)) to compute
the decorrelation kernel.

TODO DM-23857 As part of the spatially varying correction implementation
consider whether returning a Struct is still necessary.

Definition at line 118 of file imageDecorrelation.py.

             templateMatched=True, preConvMode=False, **kwargs):
         """Perform decorrelation of an image difference or of a score difference exposure.
  
         Corrects the difference or score image due to the convolution of the
         templateExposure with the A&L PSF matching kernel.
         See [DMTN-021, Equation 1](http://dmtn-021.lsst.io/#equation-1) and
         [DMTN-179](http://dmtn-179.lsst.io/) for details.
  
         Parameters
         ----------
         scienceExposure : `lsst.afw.image.Exposure`
             The original science exposure (before pre-convolution, if ``preConvMode==True``).
         templateExposure : `lsst.afw.image.Exposure`
             The original template exposure warped into the science exposure dimensions.
         subtractedExposure : `lsst.afw.image.Exposure`
             the subtracted exposure produced by
             `ip_diffim.ImagePsfMatchTask.subtractExposures()`. The `subtractedExposure` must
             inherit its PSF from `exposure`, see notes below.
         psfMatchingKernel : `lsst.afw.detection.Psf`
             An (optionally spatially-varying) PSF matching kernel produced
             by `ip_diffim.ImagePsfMatchTask.subtractExposures()`.
         preConvKernel : `lsst.afw.math.Kernel`, optional
             If not `None`, then the `scienceExposure` was pre-convolved with (the reflection of)
             this kernel. Must be normalized to sum to 1.
             Allowed only if ``templateMatched==True`` and ``preConvMode==True``.
             Defaults to the PSF of the science exposure at the image center.
         xcen : `float`, optional
             X-pixel coordinate to use for computing constant matching kernel to use
             If `None` (default), then use the center of the image.
         ycen : `float`, optional
             Y-pixel coordinate to use for computing constant matching kernel to use
             If `None` (default), then use the center of the image.
         svar : `float`, optional
             Image variance for science image
             If `None` (default) then compute the variance over the entire input science image.
         tvar : `float`, optional
             Image variance for template image
             If `None` (default) then compute the variance over the entire input template image.
         templateMatched : `bool`, optional
             If True, the template exposure was matched (convolved) to the science exposure.
             See also notes below.
         preConvMode : `bool`, optional
             If True, ``subtractedExposure`` is assumed to be a likelihood difference image
             and will be noise corrected as a likelihood image.
         **kwargs
             Additional keyword arguments propagated from DecorrelateALKernelSpatialTask.
  
         Returns
         -------
         result : `lsst.pipe.base.Struct`
             - ``correctedExposure`` : the decorrelated diffim
  
         Notes
         -----
         If ``preConvMode==True``, ``subtractedExposure`` is assumed to be a
         score image and the noise correction for likelihood images
         is applied. The resulting image is an optimal detection likelihood image
         when the templateExposure has noise. (See DMTN-179) If ``preConvKernel`` is
         not specified, the PSF of ``scienceExposure`` is assumed as pre-convolution kernel.
  
         The ``subtractedExposure`` is NOT updated. The returned ``correctedExposure``
         has an updated but spatially fixed PSF. It is calculated as the center of
         image PSF corrected by the center of image matching kernel.
  
         If ``templateMatched==True``, the templateExposure was matched (convolved)
         to the ``scienceExposure`` by ``psfMatchingKernel``. Otherwise the ``scienceExposure``
         was matched (convolved) by ``psfMatchingKernel``.
  
         This task discards the variance plane of ``subtractedExposure`` and re-computes
         it from the variance planes of ``scienceExposure`` and ``templateExposure``.
         The image plane of ``subtractedExposure`` must be at the photometric level
         set by the AL PSF matching in `ImagePsfMatchTask.subtractExposures`.
         The assumptions about the photometric level are controlled by the
         `templateMatched` option in this task.
  
         Here we currently convert a spatially-varying matching kernel into a constant kernel,
         just by computing it at the center of the image (tickets DM-6243, DM-6244).
  
         We are also using a constant accross-the-image measure of sigma (sqrt(variance)) to compute
         the decorrelation kernel.
  
         TODO DM-23857 As part of the spatially varying correction implementation
         consider whether returning a Struct is still necessary.
         """
         if preConvKernel is not None and not (templateMatched and preConvMode):
             raise ValueError("Pre-convolution kernel is allowed only if "
                              "preConvMode==True and templateMatched==True.")
  
         spatialKernel = psfMatchingKernel
         kimg = afwImage.ImageD(spatialKernel.getDimensions())
         bbox = subtractedExposure.getBBox()
         if xcen is None:
             xcen = (bbox.getBeginX() + bbox.getEndX()) / 2.
         if ycen is None:
             ycen = (bbox.getBeginY() + bbox.getEndY()) / 2.
         self.log.info("Using matching kernel computed at (%d, %d)", xcen, ycen)
         spatialKernel.computeImage(kimg, False, xcen, ycen)
  
         preConvImg = None
         if preConvMode:
             if preConvKernel is None:
                 preConvKernel = scienceExposure.getPsf().getLocalKernel()  # at average position
             preConvImg = afwImage.ImageD(preConvKernel.getDimensions())
             preConvKernel.computeImage(preConvImg, True)
  
         if svar is None:
             svar = self.computeVarianceMean(scienceExposure)
         if tvar is None:
             tvar = self.computeVarianceMean(templateExposure)
         self.log.info("Original variance plane means. Science:%.5e, warped template:%.5e)",
                       svar, tvar)
  
         if templateMatched:
             # Regular subtraction, we convolved the template
             self.log.info("Decorrelation after template image convolution")
             expVar = svar
             matchedVar = tvar
             exposure = scienceExposure
             matchedExposure = templateExposure
         else:
             # We convolved the science image
             self.log.info("Decorrelation after science image convolution")
             expVar = tvar
             matchedVar = svar
             exposure = templateExposure
             matchedExposure = scienceExposure
  
         # Should not happen unless entire image has been masked, which could happen
         # if this is a small subimage of the main exposure. In this case, just return a full NaN
         # exposure
         if np.isnan(expVar) or np.isnan(matchedVar):
             # Double check that one of the exposures is all NaNs
             if (np.all(np.isnan(exposure.image.array))
                     or np.all(np.isnan(matchedExposure.image.array))):
                 self.log.warning('Template or science image is entirely NaNs: skipping decorrelation.')
                 outExposure = subtractedExposure.clone()
                 return pipeBase.Struct(correctedExposure=outExposure, )
  
         # The maximal correction value converges to sqrt(matchedVar/expVar).
         # Correction divergence warning if the correction exceeds 4 orders of magnitude.
         mOverExpVar = matchedVar/expVar
         if mOverExpVar > 1e8:
             self.log.warning("Diverging correction: matched image variance is "
                              " much larger than the unconvolved one's"
                              ", matchedVar/expVar:%.2e", mOverExpVar)
  
         oldVarMean = self.computeVarianceMean(subtractedExposure)
         self.log.info("Variance plane mean of uncorrected diffim: %f", oldVarMean)
  
         kArr = kimg.array
         diffExpArr = subtractedExposure.image.array
         psfImg = subtractedExposure.getPsf().computeKernelImage(geom.Point2D(xcen, ycen))
         psfDim = psfImg.getDimensions()
         psfArr = psfImg.array
  
         # Determine the common shape
         kSum = np.sum(kArr)
         kSumSq = kSum*kSum
         self.log.debug("Matching kernel sum: %.3e", kSum)
  
         if preConvMode:
             self.log.info("Decorrelation of likelihood image")
             self.computeCommonShape(preConvImg.array.shape, kArr.shape,
                                     psfArr.shape, diffExpArr.shape)
             corr = self.computeScoreCorrection(kArr, expVar, matchedVar, preConvImg.array)
         else:
             self.log.info("Decorrelation of difference image")
             self.computeCommonShape(kArr.shape, psfArr.shape, diffExpArr.shape)
             corr = self.computeDiffimCorrection(kArr, expVar, matchedVar)
  
         diffExpArr = self.computeCorrectedImage(corr.corrft, diffExpArr)
  
         corrPsfArr = self.computeCorrectedDiffimPsf(corr.corrft, psfArr)
         psfcI = afwImage.ImageD(psfDim)
         psfcI.array = corrPsfArr
         psfcK = afwMath.FixedKernel(psfcI)
         psfNew = measAlg.KernelPsf(psfcK)
  
         correctedExposure = subtractedExposure.clone()
         correctedExposure.image.array[...] = diffExpArr  # Allow for numpy type casting
         # The subtracted exposure variance plane is already correlated, we cannot propagate
         # it through another convolution; instead we need to use the uncorrelated originals
         # The whitening should scale it to expVar + matchedVar on average
         if self.config.completeVarPlanePropagation:
             self.log.debug("Using full variance plane calculation in decorrelation")
             newVarArr = self.calculateVariancePlane(
                 exposure.variance.array, matchedExposure.variance.array,
                 expVar, matchedVar, corr.cnft, corr.crft)
         else:
             self.log.debug("Using estimated variance plane calculation in decorrelation")
             newVarArr = self.estimateVariancePlane(
                 exposure.variance.array, matchedExposure.variance.array,
                 corr.cnft, corr.crft)
  
         corrExpVarArr = correctedExposure.variance.array
         corrExpVarArr[...] = newVarArr  # Allow for numpy type casting
  
         if not templateMatched:
             # ImagePsfMatch.subtractExposures re-scales the difference in
             # the science image convolution mode
             corrExpVarArr /= kSumSq
         correctedExposure.setPsf(psfNew)
  
         newVarMean = self.computeVarianceMean(correctedExposure)
         self.log.info("Variance plane mean of corrected diffim: %.5e", newVarMean)
  
         # TODO DM-23857 As part of the spatially varying correction implementation
         # consider whether returning a Struct is still necessary.
         return pipeBase.Struct(correctedExposure=correctedExposure, )
  

Member Data Documentation

◆ ConfigClass

lsst.ip.diffim.imageDecorrelation.DecorrelateALKernelTask.ConfigClass = DecorrelateALKernelConfig

static

Definition at line 90 of file imageDecorrelation.py.

◆ freqSpaceShape

lsst.ip.diffim.imageDecorrelation.DecorrelateALKernelTask.freqSpaceShape

Definition at line 359 of file imageDecorrelation.py.

◆ statsControl

lsst.ip.diffim.imageDecorrelation.DecorrelateALKernelTask.statsControl

Definition at line 105 of file imageDecorrelation.py.

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

/j/snowflake/release/lsstsw/stack/lsst-scipipe-0.7.0/Linux64/ip_diffim/22.0.1-24-g1ad7a390+a9625a72a8/python/lsst/ip/diffim/imageDecorrelation.py

Public Member Functions

Static Public Member Functions

Public Attributes

Static Public Attributes

Detailed Description

Constructor & Destructor Documentation

◆ __init__()

Member Function Documentation

◆ calculateVariancePlane()

◆ computeCommonShape()

◆ computeCorrectedDiffimPsf()

◆ computeCorrectedImage()

◆ computeDiffimCorrection()

◆ computeScoreCorrection()

◆ computeVarianceMean()

◆ estimateVariancePlane()

◆ padCenterOriginArray()

◆ run()

Member Data Documentation

◆ ConfigClass

◆ freqSpaceShape

◆ statsControl

◆ init()