Comparison with CTA-MARS: Direction reconstruction

[HealthyPear]

This issue is part of a project described in issue #24.

The following is a "real-time" list of points that are found to be differences between the pipelines using the comparison.
Not all features are critical to recovering the missing performance, but all should be implemented (as more similar as possible) in order to allow their optional use when comparing different algorithms.

---

• Basic quality cuts (configurable)

• Intensity>50 (in the "biased" units - see Comparison with CTA-MARS: Calibration #31 and the calibration benchmarking notebook)

• 0.1 < Width / Length < 0.6

• 80% camera radius containment

• Direction LUT and associated quality cut

From CTAMARS wiki page:

"a look-up table is built per telescope type, containing, in bins of image intensity and Width / Length, the square of . This allows us to estimate for an image, given its Intensity, Width and Length, how reliable its axis is as a measure of the shower axis' orientation. The values in these LUTs will be used to set relative weights for the different telescopes i the stereoscopic reconstruction of events with three or more valid images."

The associated image quality cut is number of events in the corresponding direction LUT bin <10.

• Reconstructor

From IRF Prod3b report,

"A maximum likelihood fit is performed to obtain the shower axis parameters (direction, core location).
In this fit each image takes part with a weight calculated according to its elongation (Width/Length) and
Size (total light content)."

This could be implemented in ctapipe (see cta-observatory/ctapipe#1636)

The weight mentioned here comes from the Direction LUT.