search

DUNE / dune-tms

DUNE ND Temporary Muon Spectrometer
0 stars 1 forks source link

Many missing hits in the tracks #45

Closed AsaNehm closed 9 months ago

AsaNehm commented 9 months ago

In a lot of tracks, there are a lot of missing hits. Meaning there are hits that should be part of a track as they are directly on the Hough line, but they are not part of the track. spill_040

AsaNehm commented 9 months ago

This issue seems to appear the most in the presence of multiple tracks. So looking into the option of merging tracks after the Hough transform (config/TMS_Default... [Recon.Hough] MergeTracks)

AsaNehm commented 9 months ago

At some point the newly merged hits into the previous track get left out. This seems to appear during the A attempt at the last step of the Hough transform. (edit) It seems that the A is not connecting the merged tracks but rather running on them separately and then only really pushing back the unmerged longer one

AsaNehm commented 9 months ago

The missing hits appear to have the same plane numbers as the hits that are already connected in the track and therefore get ignored. The crux: the z coordinates are actually not the same, and they should actually be in a different plane -> gaps in track as draw_spill uses coordinates not plane/bar number for plotting. But how to correct this problem?

AsaNehm commented 9 months ago

There might also only be 50 planes implemented instead of the actual 100!!! (z coordinate 17833 -> plane number 46)

AsaNehm commented 9 months ago

Geometry implementation has for each orientation type 50 layers and numbers them in the same orientation type consecutively. Meaning two layers can have the same plane number if they are in different orientation types. Fix is to make the reconstruction also name sensitive and adapting everything also to more than just X/YBars as current orientation types are Y and V

AsaNehm commented 9 months ago

TMS_Nehm_002_code.pdf