private void PairSkyBot(List <Tracklet> Tracklets, double ArcLengthSec, string ReportFieldName, int CCDNumber, FitsImage[] ImageSet, Step.StepPipeline Pipeline)
{
foreach (Image img in ImageSet)
{
SkyBotImageData skid = img.GetProperty <SkyBotImageData>();
//skid.RetrieveObjects(ObservatoryCode);
foreach (Tracklet tk in Tracklets)
{
skid.TryPair(tk, ArcLengthSec);
}
var unp = skid.GetUnpaired();
if (unp.Count != 0)
{
Logger("SkyBoT: Unpaired objects left: ");
foreach (var o in unp)
{
string SkData = "SkyBoT: " + ExtraIO.EquatorialPointStringFormatter.FormatToString(o, ExtraIO.EquatorialPointStringFormatter.Format.MPC);
PixelPoint pp = img.Transform.GetPixelPoint(o);
SkData += ";" + pp.ToString();
Logger(SkData);
var Reasons = Pipeline.QueryWhyNot(o, 5);
string LRs = "Reasons: ";
foreach (string x in Reasons)
{
LRs += x + ";";
}
if (Reasons.Count == 0)
{
Logger("Not found in removal log, so not detected.");
}
else
{
Logger(LRs);
}
}
}
}
for (int i = 0; i < Tracklets.Count; i++)
{
Tracklet tk = Tracklets[i];
if (!tk.TryFetchProperty(out ObjectIdentity objid))
{
objid = new ObjectIdentity();
}
objid.ComputeNamescoreWithDefault(tk, null, ReportFieldName, CCDNumber, i);
tk.SetResetProperty(objid);
}
}
public List <Tracklet> AnalyzeCCD(PipelineArguments Args)
{
Logger("Setting up pipeline");
/* Deal with incorrect SWARP flux scaling */
SWarpScaling.ApplyTransform = CorrectSWARP;
string RunDir = Args.RunDir;
if (!Directory.Exists(RunDir))
{
Directory.CreateDirectory(RunDir);
}
/* Read input images and preprocess for poisson noise */
int ImageCount = Args.Inputs.Length;
FitsImage[] FirstProcess = new FitsImage[ImageCount];
double[] PFW = PipelineHelperFunctions.LinearizedPoissonKernel(PoissonRadius);
Step.StepPipeline sp = new Step.StepPipeline(StandardBITPIX, RunDir, Args.Inputs.Length, MaxDetections);
sp.LogHookImage = LogImage;
sp.LogHookDetection = LogDet;
sp.LogMessage = LogMessage;
bool HasBadpix = Args.Badpixel != null;
Logger("Begining to run the pipeline");
var zpTask = System.Threading.Tasks.Task <Dictionary <IO.Image, double> > .Factory.StartNew(() => CalibrateZP(Args.Inputs));
var skTask = System.Threading.Tasks.Task <bool> .Factory.StartNew(() => PrecacheSkyBot(Args.Inputs));
BitArray[] map = PipelineHelperFunctions.ExtractBadpixel(Args.Badpixel, Logger);
for (int i = 0; i < ImageCount; i++)
{
FitsImage Pipeline = Args.Inputs[i];
Pipeline.GetProperty <ImageSource>().AddToSet(Pipeline, "Original");
sp.SetModel(i, Pipeline, new List <IO.ImageProperties>()
{
Pipeline.GetProperty <ObservationTime>()
});
if (!UseCoreFilter)
{
sp.RunPipeline(RestrictedMean.RestrictedMeanFilter, "Poisson", i, ref Pipeline, PFW, RestrictedMean.Parameters(PoissonRadius));
}
else if (HasBadpix)
{
sp.RunPipeline(CoreFilter.Filter, "Poisson", i, ref Pipeline, new CoreFilter.CoreFilterParameters(PFW, map), CoreFilter.Parameters(PoissonRadius));
}
else
{
throw new ArgumentException("Must specify Badpixel files if trying to run with CoreFilter");
}
if (Operations.HasFlag(EnabledOperations.Normalization))
{
if (!sp.EnsureImage("Normalized", i, out FitsImage Normalized))
{
Point4Distance p4d = new Point4Distance(Pipeline, Normalized, NormalizationMeshSize);
Logger("Generated Normalized image " + i);
}
else
{
Logger("Found Normalized image " + i);
}
Normalized.GetProperty <ImageSource>().AddToSet(Args.Inputs[i], "Normalized");
Pipeline = Normalized;
}
FirstProcess[i] = Pipeline;
}
/* Create the central median */
string CentralPath = Path.Combine(RunDir, "Central.fits");
if (!sp.EnsureCentralImage("Central", out FitsImage Central))
{
HardMedians.MultiImageMedian.Run(null, FirstProcess, Central, HardMedians.MultiImageMedianParameters);
Logger("Generated Central image");
}
else
{
Logger("Found Central image");
}
Logger("Computed the multi-image median");
/* Prepare the mask, slow object detector, trail detector, weights for second median filtering, etc. */
ImageStatistics CentralStats = new ImageStatistics(Central);
if (Args.Clipped)
{
CentralStats = new ImageStatistics(Central, CentralStats.ZeroLevel, 2 * CentralStats.StDev);
}
StarData StarList = new StarData();
ComputeDetectorData(Central, CentralStats, StarList, out MaskByMedian.MaskProperties MaskProp, out DotDetector SlowDetector,
out LongTrailDetector.LongTrailData LTD);
if (Args.Clipped)
{
SlowDetector.HighThresholdMultiplier *= 2;
SlowDetector.LowThresholdMultiplier *= 2;
}
DetectionReducer dr = new DetectionReducer()
{
PairingRadius = 0.7
};
if (Operations.HasFlag(EnabledOperations.SourceExtractor))
{
try
{
dr.LoadStars(StarList.FixedStarList);
}
catch (Exception ex) { throw new ArgumentException("Could not read detections from SE catalog.", ex); }
dr.GeneratePool();
}
Logger("Set up detectors");
List <ImageDetection> FullDetectionsList = new List <ImageDetection>();
double[] FMW2 = PipelineHelperFunctions.LinearizedMedianKernel();
LTLimit ltl = new LTLimit()
{
MinPix = TrailMinPix
};
RipFilter rf = new RipFilter()
{
SigmaTop = 30
};
Logger("Ready for final image processing and detection");
for (int i = 0; i < ImageCount; i++)
{
List <ImageDetection> LocalDetectionList = new List <ImageDetection>();
FitsImage DetectionSource = FirstProcess[i];
if (Operations.HasFlag(EnabledOperations.Masking))
{
sp.RunPipeline(MaskByMedian.Masker, "Masked", i, ref DetectionSource, MaskProp, MaskByMedian.Parameters);
}
if (Operations.HasFlag(EnabledOperations.SecondMedian))
{
sp.RunPipeline(HardMedians.WeightedMedian, "Second Median", i, ref DetectionSource, FMW2, HardMedians.WeightedMedianParameters(SecMedRadius));
}
ImageStatistics SecMedStat = new ImageStatistics(DetectionSource);
if (Operations.HasFlag(EnabledOperations.LongTrailDetector))
{
var Dets = sp.RunDetector((FitsImage img) =>
{
LongTrailDetector.PrepareAlgorithmForImage(img, SecMedStat, ref LTD);
LongTrailDetector.Algorithm.Run(LTD, DetectionSource, LongTrailDetector.Parameters);
return(LTD.Results);
}, DetectionSource, "Trail", DetectionAlgorithm.Trail);
LocalDetectionList.AddRange(Dets);
}
if (Operations.HasFlag(EnabledOperations.BlobDetector))
{
var Dets = sp.RunDetector(SlowDetector.Detect, DetectionSource, "Blob", DetectionAlgorithm.Blob);
LocalDetectionList.AddRange(Dets);
}
if (Operations.HasFlag(EnabledOperations.SourceExtractor))
{
var dts = sp.RunDetector((arg) =>
{
List <ImageDetection> Dets = ExtraIO.SourceExtractor.ParseSEFile(Args.CatalogData[i], Args.Inputs[i]);
Dets = Dets.Where((x) => x.FetchProperty <ObjectPhotometry>().Flux > 300).ToList();
var ND = dr.Reduce(Dets);
return(ND);
}, DetectionSource, "SE", DetectionAlgorithm.SourceExtractor);
LocalDetectionList.AddRange(dts);
}
if (Operations.HasFlag(EnabledOperations.OutputDetectionMap))
{
DetectionDebugMap(RunDir, i, LocalDetectionList, DetectionSource);
}
rf.ImgMean = SecMedStat.ZeroLevel;
rf.ImgSigma = SecMedStat.StDev;
var NLDL = sp.RunFilters(LocalDetectionList, "LocalToGlobal", ltl, rf);
Logger("Total " + NLDL.Count + " detections.");
FullDetectionsList.AddRange(NLDL);
}
Logger("Filtering and pairing detections...");
LinearityThresholdFilter LTF = new LinearityThresholdFilter()
{
MaxLineThickness = MaxLineThickness
};
List <ImageDetection> FilteredDetections = sp.RunFilters(FullDetectionsList, "MainFilter", LTF);
StarList.MarkStarCrossed(FilteredDetections, StarCrossRadiusM, StarCrossMinFlux);
if (Args.CCDBadzone != null)
{
FilteredDetections = sp.RunFilters(FilteredDetections, "Badzone", Args.CCDBadzone);
}
Logger("Before PrePair " + FilteredDetections.Count);
PrePair.MatchDetections(FilteredDetections, MaxPairmatchDistance, MixMatch, SameArcSep);
Logger("Left with " + FilteredDetections.Count + " detections");
LinePoolSimple lps = new LinePoolSimple()
{
MaxLinErrorArcSec = MaxResidual, SearchExtraSmall = SmallExtraSearchRadius, SearchExtraBig = BigExtraSearchRadius
};
lps.LoadDetections(FilteredDetections);
lps.GeneratePool();
var Pairings = lps.FindTracklets();
sp.NotePairings(FilteredDetections, Pairings);
Logger("Found " + Pairings.Count + " raw tracklets");
LinearityTest lintest = new LinearityTest();
StaticFilter stf = new StaticFilter();
TotalError te = new TotalError();
var TK2List = sp.RunFilters(Pairings, "Tracklet Filtering", stf, te);
Logger("After filtering: " + TK2List.Count + " candidate objects found");
sp.LogDetections(Path.Combine(RunDir, "detlog.txt"));
Dictionary <IO.Image, double> ZP = zpTask.Result;
skTask.Wait();
var Recovered = RecoverTracklets(TK2List, Args.Inputs, Path.Combine(RunDir, "reclog.txt"), ZP);
TrackletsDeduplication.Deduplicate(Recovered, 1.0);
Logger("Recovered " + Recovered.Count + " candidate objects");
PairSkyBot(Recovered, SkyBoTDistance, Args.FieldName, Args.CCDNumber, Args.Inputs, sp);
return(Recovered);
}