Intensity Filter Widget

The Intensity Filter widget allows users to down-select diffraction peaks based on their scattering strength.

Functionality

This widget allows you to:

1. Normalize Intensity: Calculate a filtered intensity metric by combining raw sum intensity with a reciprocal distance (\(ds\)​) correction.

2. Threshold Peaks: Apply a fractional cutoff to keep only the most significant reflections.

3. Visualize the Cutoff: Compare the distribution of all peaks against those that pass the intensity filter in a real-time plot.

User Interface

Control Panel

These two parameters work together to define which peaks are “strong enough” to be kept:

1. Intensity Fraction (intensity_frac): A percentile-based threshold (between 0.0 and 1.0). It defines the global cutoff for the normalized intensity metric.

   • Example: A value of 0.95 will filter out the weakest peaks, retaining only the top 5% of the normalized signal.

2. Thermal Factor: This parameter adjusts the filtering strength relative to the reciprocal distance (\(ds\)​).

   • Low values: Treat peaks at low and high \(ds\)​ similarly.

   • High values: Increase the “penalty” for peaks at high \(ds​\), effectively allowing you to filter weak intensity more aggressively at lower rings while being more “mild” or permissive at higher rings where signal naturally weakens.

Execution Control

• Process Group Name: The specific HDF5/Nexus path where the filtered peak data will be stored

  1. defaults to lattice_filtered_peaks for Lattice filtering.

  2. defaults to intensity_filtered_peaks for Intensity filtering.

  3. defaults to indexer_filtered_peaks for Indexer filtering.

  Tip:

  By assigning a unique name here (e.g., alpha_phase_peaks or beta_phase_peaks), you can process multiple phases within the same data file. This allows you to run independent Indexing and Grain Mapping tasks for each phase subsequently.

• Filter Peaks: Executes the filtering.

• Overwrite: If checked, existing groups with the same process name will be replaced.

Example Usage

1. Connect or Load: Connect the widget to a Lattice Filter output or manually browse to a Nexus file containing a filtered data group.

2. Initial Observation: Observe the blue distribution (unfiltered peaks) in the scatter plot.

3. Set Global Cutoff: Adjust the Intensity Fraction to set your baseline noise floor (e.g., 0.98).

4. Refine with Thermal Factor: Adjust the Thermal Factor to fine-tune the filter across the rings. Use higher values if you find that you are losing too many valid peaks at high ds​ compared to the low \(ds\)​ region.

5. Execute: Click Filter Peaks. The red points will indicate the peaks that passed the filter.