Skip to content

v0.15.0
Compare
Choose a tag to compare
@kbruegge kbruegge released this 02 Jun 15:31
· 1232 commits to master since this release
v0.15.0
73692e2

Version 0.15.0 -- 23.05.2016

  • Muon Analysis
    • New Processor RingStandardDeviationWithThreshold to calculate Standard Deviation of the ArrivalTime on MyonRing pixels above a threshold.
    • Update example xml for muon study (examples/studies/muon_identification.xml)
    • fact.features.muon.HoughTransform now uses a PixelSet for the BestRingPixels
    • Default outputkeys of fact.features.muon.GaussianFit and fact.features.muon.CircularFit now use CamelCase instead of snake_case.
  • Improved error handling in source position calculation and AuxServices. The process now properly stops in case an IO related error occurs.
  • New Features: Watershed
    • New package Watershed, containing ClusterFellwalker, ClusterArrivalTimes, FactCluster
    • ClusterFellwalker (Processor) creates new features based on a watershed algorithm called FellWalker. This algorithm groups camera pixels by their values for photoncharge or something else. From this clusters, new features based on number,
      content, shape and position in camera are extracted.
    • ClusterArrivalTimes (Processor) also creates new features, based on another algorithm (Region Growing). In this case, pixels are grouped by their arrival times. The extracted features are similar to ClusterFellwalker.
    • FactCluster (Object) holds all features and information about a cluster; used by ClusterFellwalker and ClusterArrivaltimes
  • MotionCleaning
    • Idea for a cleaning: Select all pixels which show a strong increase in brightness over the time series in a time window from slice 50 to slice 120.
      This cleaning does not select shower as compact and exactly as standard cleanings!!! There are many more islands surviving the cleaning. Could be useful, but not for the standard analysis settings at the moment!!
  • CleaningPerformance
    • Calculates performance values as precision, recall, accuracy true/false positive/negative for a cleaning.
  • SmoothBell
    • Processor to smooths a camera image (not the time series of a pixel). Calculates the mean of six times the value of a pixel plus values of all neighbor pixels.
  • MeanCorrelation
    • Bug fix: no more errors due to division by zero in case the time series contains zero.
Assets 3
Loading