Sorting out requirements from a high-level requirement note of ODIN

[YooSunYoung]

I just got a long list of what should be done for ODIN.
It has requirements for all software stacks not only data-reduction so we should cherry-pick relevant requirements for imaging.
It also has requirements related to diffraction.

List of High-level Requirements for ODIN.

File IO

• Read/write different formats of images [Requirement] [File-Handling] Support various image formats. #34
• Pre-defined meta data formats
• Loading sub-parameters separately, i.e. sample environment parameters
• Splitting file into different measurement modes. [Requirement] [File-Handling] Splitting a single file into multiple measurement modes. #35
• McStas file loading in a format that can be used in a workflow.

WFM

• Verify resolution (fitting Bragg edge width w. calibration sample “Felix”)
  Selection of delta_lambda/lambda for adjusting position of WFM1 and WFM2
• Selection of Bandwidth (Adjusted WFM positions)
  Verification (“Felix”, “Beatrix”) of visibility and position of Bragg edges

Optics and spatial informations

• Field of view/resolution in cameras (incl. various scintillator screens), i.e. Siemens star: [Requirement] SiemensStar spatial resolution measurement #17

• Position of detectors (Downstream and perpendicular to the sample position):
  Odin has a positioning system (300 microns precision)
  Se Geometry in Calibration section below

Acquisition modalities

• General acquisition modes: Radiography, Step by step, Continuous, Golden Ratio
  In combination with steering of SE
• Attenuation Tomography
• Diffraction (Transmission): Bragg edge (and transmission Grain mapping)
• Diffraction (Perpendicular): Powder, Texture, Grain mapping
• Gratings
  Horizontal translation of gratings (and rotation for alignment of gratings)
• Polarized neutrons: Radiography, Tensor based tomography, Spin-Phase option
• Stroboscopic: imaging of fast repetitive processes (Re-sorting and re-naming (stroboscopic – phase – time sorting[MOU1] ..) process not in sync with source
  We can use the tdct(?)
  We have to define the time-window.
• Exposure: Criteria on acquisition length??

Calibration

Efficiency: (both transmission and diffraction mode detectors) Sample: Vanessa

• Vanadium + normalization
• Beam monitor as normalization (and proton current): also, HC step.

Geometry

• Use positioning system
• Sample to detector distance (transmission)
  Bragg edge fitting on Fe sample (Matlab -> python)
• Detector distance + tilt (diffraction)
  Indexing and constrained fit to Single crystal, polycrystal and powder

Others

• SE (without sample) -> constitutes the open beam
• Gratings: To be specified

Events to Images

• White beam integration (rebin spatial resolution)
• BC: Binning to images (x,y,t)
• WFM stitching (both in event mode and binned data)

Normalization

• Specify BB (not ready yet), OB, DC
• Normalization based on a ROI (like brightness adjustment on a camera)

Reconstruction (+ available software)

• Attenuation tomography. i.e. pyMuhrec, backup Savu??
• Bragg edge: ToFLib (GUI work from SNS, Jean B)
• Grains in transmission mode: Trindex
• Gratings: TaPy
• Polarized neutrons: Should be implemented
• Diffraction: 6DNXRD, BEER software
• Rietveld refinement, Transmission: RITS, BeTmen (Sven) - data analysis part
• Rietveld refinement, Diffraction: FullProf - data analysis part

Simulation

• McStas data for comparison (SE and their components as unions)

Data Reduction

• WFM stitching (as 1D and 2D) as option before AND/OR after the normalization/analysis routines.
• Comparison of the two cases:
  - Binning full spectrum into images + Normalization + Stitching.
  - Stitching events + Normalization
• Simple conversion between TOF and lambda
• Conversion between attenuation coefficients and transmission
  :math:func(attenuation_coefficiency, sample_thickness) = transmission
  func is based on Beer Lambert Law
• Black-body correction (developed by PSI, correction according to transmission)
• Fitting routines i.e. siemens stars spatial resolution

Image Tools

Basically making image J and Fiji
We should check if Fiji provides APIs. (Still not fast enough)

• Simple mathematical operations: scaling, adding,…
• filtering, averaging (in all dimensions)
• histograms of ROIs
• data extraction from ROIs, line plots, z-plots
• binning (2D and 3D)
• ROI flexibility: specifically very flexible ROI tools (more than one, generic shapes)
• ROI’s in tof/wavelength
• complex ROI shape (free hand in 3D).
• visualization tools (contrast, brightness…) available at selected processing step
• Re-sorting and re-naming (stroboscopic – phase – time sorting[MOU2] ..) process not in sync with source
• Additional: running averages and interpolation tools in x-y-lambda-t
  We have to discuss we should state it clearly.
• Spatial image stitching (also workable for tomography) – check VGStudio max or similar, Amira/Avizo
• extraction (+plotting) of TOF dependency (e.g. Bragg edge patterns) from Voxels
• Display/Visualization (images) of refinement parameters (including uncertainties) for quality check.
• Convenient access to fitted spectra of each pixel.
• Rodrigues vectors and quaternions
• Visualization tools(contrast, brightness, etc) - we can use external tools maybe?
• Target tracking tool

[YooSunYoung]

We're gonna separate these based on user stories.

[YooSunYoung]

I'll remove it from the board since it's all collected here.