skyObjects.py

Go to the documentation of this file.

 
__all__ = ["SkyObjectsConfig", "SkyObjectsTask", "generateSkyObjects"]
 
from scipy.stats import qmc
 
from lsst.pex.config import Config, Field, ListField
from lsst.pipe.base import Task
 
import lsst.afw.detection
import lsst.afw.geom
import lsst.afw.math
 
 
class SkyObjectsConfig(Config):
    """Configuration for generating sky objects"""
    avoidMask = ListField(dtype=str, default=["DETECTED", "DETECTED_NEGATIVE", "BAD", "NO_DATA"],
                          doc="Avoid pixels masked with these mask planes")
    growMask = Field(dtype=int, default=0,
                     doc="Number of pixels to grow the masked pixels when adding sky objects")
    sourceRadius = Field(dtype=float, default=8, doc="Radius, in pixels, of sky objects")
    nSources = Field(dtype=int, default=100, doc="Try to add this many sky objects")
    nTrialSources = Field(dtype=int, default=None, optional=True,
                          doc="Maximum number of trial sky object positions "
                              "(default: nSkySources*nTrialSkySourcesMultiplier)")
    nTrialSourcesMultiplier = Field(dtype=int, default=5,
                                    doc="Set nTrialSkySources to "
                                        "nSkySources*nTrialSkySourcesMultiplier "
                                        "if nTrialSkySources is None")
 
 
def generateSkyObjects(mask, seed, config):
    """Generate a list of Footprints of sky objects
 
    Sky objects don't overlap with other objects. This is determined
    through the provided `mask` (in which objects are typically flagged
    as `DETECTED`).
 
    Sky objects are positioned using a quasi-random Halton sequence number
    generator. This is a deterministic sequence that mimics a random trial and
    error approach whilst acting to minimize clustering of points for a given
    field of view. Up to `nTrialSources` points are generated, returning the
    first `nSources` that do not overlap with the mask.
 
    Parameters
    ----------
    mask : `lsst.afw.image.Mask`
        Input mask plane, which identifies pixels to avoid for the sky
        objects.
    seed : `int`
        Random number generator seed.
    config : `SkyObjectsConfig`
        Configuration for finding sky objects.
 
    Returns
    -------
    skyFootprints : `list` of `lsst.afw.detection.Footprint`
        Footprints of sky objects. Each will have a peak at the center
        of the sky object.
    """
    if config.nSources <= 0:
        return []
 
    skySourceRadius = config.sourceRadius
    nSkySources = config.nSources
    nTrialSkySources = config.nTrialSources
    if nTrialSkySources is None:
        nTrialSkySources = config.nTrialSourcesMultiplier*nSkySources
 
    box = mask.getBBox()
    box.grow(-(int(skySourceRadius) + 1))  # Avoid objects partially off the image
    xMin, yMin = box.getMin()
    xMax, yMax = box.getMax()
 
    avoid = lsst.afw.geom.SpanSet.fromMask(mask, mask.getPlaneBitMask(config.avoidMask))
    if config.growMask > 0:
        avoid = avoid.dilated(config.growMask)
 
    sampler = qmc.Halton(d=2, seed=seed).random(nTrialSkySources)
    sample = qmc.scale(sampler, [xMin, yMin], [xMax, yMax])
 
    skyFootprints = []
    for x, y in zip(sample[:, 0].astype(int), sample[:, 1].astype(int)):
        if len(skyFootprints) == nSkySources:
            break
 
        spans = lsst.afw.geom.SpanSet.fromShape(int(skySourceRadius), offset=(x, y))
        if spans.overlaps(avoid):
            continue
 
        fp = lsst.afw.detection.Footprint(spans, mask.getBBox())
        fp.addPeak(x, y, 0)
        skyFootprints.append(fp)
 
        # Add doubled-in-size sky object spanSet to the avoid mask.
        avoid = avoid.union(spans.dilated(int(skySourceRadius)))
 
    return skyFootprints
 
 
class SkyObjectsTask(Task):
    """Generate a list of Footprints of sky objects.
    """
    ConfigClass = SkyObjectsConfig
 
    def run(self, mask, seed):
        """Generate a list of Footprints of sky objects
 
        Sky objects don't overlap with other objects. This is determined
        through the provided `mask` (in which objects are typically flagged
        as `DETECTED`).
 
        Sky objects are positioned using a quasi-random Halton sequence
        number generator. This is a deterministic sequence that mimics a random
        trial and error approach whilst acting to minimize clustering of points
        for a given field of view. Up to `nTrialSources` points are generated,
        returning the first `nSources` that do not overlap with the mask.
 
        Parameters
        ----------
        mask : `lsst.afw.image.Mask`
            Input mask plane, which identifies pixels to avoid for the sky
            objects.
        seed : `int`
            Random number generator seed.
 
        Returns
        -------
        skyFootprints : `list` of `lsst.afw.detection.Footprint`
            Footprints of sky objects. Each will have a peak at the center
            of the sky object.
        """
        skyFootprints = generateSkyObjects(mask, seed, self.config)
        self.log.info("Added %d of %d requested sky sources (%.0f%%)", len(skyFootprints),
                      self.config.nSources, 100*len(skyFootprints)/self.config.nSources)
        return skyFootprints