Validation: BRDF correction

Recorded evidence: 5 variations; 5 passed, 0 failed, and 0 skipped (100.0% pass rate over all recorded variations).

Evidence boundary

The current checked-in campaign uses small synthetic or already-present inputs and does not contact NEON. It validates software contracts and diagnostics, not real-flightline scientific accuracy.

What this module test exercises

Use a neutral BRDF model as an identity contract across view angles, storage scales, cube sizes, and band counts.

Implementation exercised: apply_brdf_correct

Inputs varied

Field	Why it is recorded
view_zenith_degrees	Varies view geometry from nadir to 28°.
scale_factor	Alternates unit and scaled reflectance storage.
shape_y_x_b	Varies spatial dimensions and band count.

Checks and how to interpret them

Check	Question	PASS means	If it does not pass
shape_preserved	Does BRDF application retain the cube shape?	Output and input dimensions match exactly.	Inspect band axis and tile assembly.
neutral_model_is_identity	Does an iso=1, vol=0, geo=0 model leave reflectance unchanged?	Input and output agree within 1e-5 stored units.	Any drift indicates a kernel, scaling, or factor-application regression.
dtype_preserved	Does correction retain the float32 output contract?	The output dtype is float32.	Review memory allocation and NumPy promotion before accepting larger files.

Diagnostics recorded for every variation

Field	Why it is recorded
max_absolute_error_stored_units	Largest identity-model difference.
finite_percent	Percent of numerically defined corrected values.
output_min_unitless	Minimum output after conversion to unit reflectance.
output_max_unitless	Maximum output after conversion to unit reflectance.

Input variations and results

On narrow screens, scroll the table horizontally to see every diagnostic and check.

Variation	Input variation	Result	Diagnostics	Explicit checks
brdf_correction-001 Neutral BRDF model at 0° view zenith.	scale_factor=1; shape_y_x_b=[5,6,2]; view_zenith_degrees=0	PASS	finite_percent=100; max_absolute_error_stored_units=0; output_max_unitless=0.626875; output_min_unitless=0.0513522	dtype_preserved=✓; neutral_model_is_identity=✓; shape_preserved=✓
brdf_correction-002 Neutral BRDF model at 7° view zenith.	scale_factor=0.0001; shape_y_x_b=[6,7,3]; view_zenith_degrees=7	PASS	finite_percent=100; max_absolute_error_stored_units=0; output_max_unitless=0.64909; output_min_unitless=0.0542199	dtype_preserved=✓; neutral_model_is_identity=✓; shape_preserved=✓
brdf_correction-003 Neutral BRDF model at 14° view zenith.	scale_factor=1; shape_y_x_b=[7,8,4]; view_zenith_degrees=14	PASS	finite_percent=100; max_absolute_error_stored_units=0; output_max_unitless=0.643089; output_min_unitless=0.0510967	dtype_preserved=✓; neutral_model_is_identity=✓; shape_preserved=✓
brdf_correction-004 Neutral BRDF model at 21° view zenith.	scale_factor=0.0001; shape_y_x_b=[8,6,5]; view_zenith_degrees=21	PASS	finite_percent=100; max_absolute_error_stored_units=0; output_max_unitless=0.649915; output_min_unitless=0.0522228	dtype_preserved=✓; neutral_model_is_identity=✓; shape_preserved=✓
brdf_correction-005 Neutral BRDF model at 28° view zenith.	scale_factor=1; shape_y_x_b=[5,7,2]; view_zenith_degrees=28	PASS	finite_percent=100; max_absolute_error_stored_units=0; output_max_unitless=0.643964; output_min_unitless=0.0507589	dtype_preserved=✓; neutral_model_is_identity=✓; shape_preserved=✓

What a passing result establishes

Neutral-model invariance and basic numerical stability.

What it does not establish

Accuracy of fitted BRDF coefficients for real angular sampling.

The matching real stage checks are explained in the stage QA test guide.

Example from the real R10C test run

The real run displays fitted BRDF profiles and unfiltered geometry summaries; four fields are marked for range review.

The figure is evidence from one completed flightline, not a replacement for the variation table above. Open the real flightline walkthrough for exact values and limitations.

Expansion to 100 real variations

The repository includes a live 100-flightline campaign specification. It requires a pinned inventory of real flightline IDs plus an explicit compute, storage, and network allocation. Live results must be stored as a new campaign record; they must not overwrite this offline baseline.

Reproduce or expand this module

# Fast local evidence matrix (five variations per module)
python scripts/run_validation_campaign.py --iterations-per-module 5

# Exercise 100 deterministic small-data variations per module
python scripts/run_validation_campaign.py --iterations-per-module 100 \
  --output validation/results/offline-contract-100.json

python scripts/generate_validation_docs.py

The 100-case offline command scales contract variation and randomized synthetic inputs. It does not substitute for 100 distinct NEON downloads.

Last updated: 2026-08-14