lsst.meas.astrom.sip.sourceMatchStatistics Namespace Reference

Functions
	sourceMatchStatistics (matchList, log=None)

Function Documentation

sourceMatchStatistics()

lsst.meas.astrom.sip.sourceMatchStatistics.sourceMatchStatistics	(		matchList,
			log = None )

Compute statistics on the accuracy of a wcs solution, using a
precomputed list of matches between an image and a catalog.

Parameters
----------
matchList : `lsst.afw.detection.SourceMatch`
    List of matches between sources and references to compute statistics
    on.

Returns
-------
values : `dict
    Value dictionary with fields:

    - diffInPixels_mean : Average distance between image and
      catalog position in pixels (`float`).
    - diffInPixels_std : Root mean square of distribution of distances
      (`float`).
    - diffInPixels_Q25 : 25% quantile boundary of the match dist
      distribution (`float`).
    - diffInPixels_Q50 : 50% quantile boundary of the match dist
      distribution (`float`).
    - diffInPixels_Q75 : 75% quantile boundary of the match
      dist distribution (`float`).

Definition at line 28 of file sourceMatchStatistics.py.

def sourceMatchStatistics(matchList, log=None):
    """Compute statistics on the accuracy of a wcs solution, using a
    precomputed list of matches between an image and a catalog.
 
    Parameters
    ----------
    matchList : `lsst.afw.detection.SourceMatch`
        List of matches between sources and references to compute statistics
        on.
 
    Returns
    -------
    values : `dict
        Value dictionary with fields:
 
        - diffInPixels_mean : Average distance between image and
          catalog position in pixels (`float`).
        - diffInPixels_std : Root mean square of distribution of distances
          (`float`).
        - diffInPixels_Q25 : 25% quantile boundary of the match dist
          distribution (`float`).
        - diffInPixels_Q50 : 50% quantile boundary of the match dist
          distribution (`float`).
        - diffInPixels_Q75 : 75% quantile boundary of the match
          dist distribution (`float`).
    """
 
    size = len(matchList)
    if size == 0:
        raise ValueError("matchList contains no elements")
 
    dist = np.zeros(size)
    i = 0
    for match in matchList:
        catObj = match.first
        srcObj = match.second
 
        cx = catObj.getXAstrom()
        cy = catObj.getYAstrom()
 
        sx = srcObj.getXAstrom()
        sy = srcObj.getYAstrom()
 
        dist[i] = np.hypot(cx-sx, cy-sy)
        i = i+1
 
    dist.sort()
 
    quartiles = []
    for f in (0.25, 0.50, 0.75):
        i = int(f*size + 0.5)
        if i >= size:
            i = size - 1
        quartiles.append(dist[i])
 
    values = {}
    values['diffInPixels_Q25'] = quartiles[0]
    values['diffInPixels_Q50'] = quartiles[1]
    values['diffInPixels_Q75'] = quartiles[2]
    values['diffInPixels_mean'] = dist.mean()
    values['diffInPixels_std'] = dist.std()
 
    return values