template<typename PixelT>
class lsst::ip::diffim::detail::KernelSumVisitor< PixelT >
A class to accumulate kernel sums across SpatialCells.
ps->set("kernelSumClipping", false);
ps->set("maxKsumSigma", 3.0);
detail::KernelSumVisitor<PixelT> kernelSumVisitor(*ps);
kernelSumVisitor.reset();
kernelCells.visitCandidates(&kernelSumVisitor, nStarPerCell);
kernelSumVisitor.processKsumDistribution();
kernelCells.visitCandidates(&kernelSumVisitor, nStarPerCell);
int nRejected = kernelSumVisitor.getNRejected();
- Note
- The class has 2 processing modes; the first AGGREGATES kernel sums across all candidates. You must the process the distribution to set member variables representing the mean and standard deviation of the kernel sums. The second mode then REJECTs candidates with kernel sums outside the acceptable range (set by the ps). It does this by setting candidate status to afwMath::SpatialCellCandidate::BAD. In this mode it also accumulates the number of candidates it sets as bad.
-
The statistics call calculates sigma-clipped values (afwMath::MEANCLIP, afwMath::STDEVCLIP)
Definition at line 27 of file KernelSumVisitor.h.