Qualification Review (QR) report

Qualification Review (QR) report#

Project: s2_msi_raw_generator · ECSS-E-ST-40C Rev.1 qualification review of the v0.3.0 baseline (2026-07-02). The qualified capability is the reverse chain: running a Sentinel-2B L1B backwards through the operational L0→L1B radiometric chain to reconstruct L1A → L0plus → Synthetic L0, validated against the reference ESA L0 img. Review implementation: SRevP (QR-equivalent held as release MR !32 + this consolidated report).

1. QR objectives#

Confirm that (a) the requirements baseline is verified, (b) the document set is complete per the tailored DRL, (c) the acceptance criteria of the S2 L1B E2E — the reverse chain — are met, i.e. the Synthetic L0 agrees with the reference ESA L0 img (10/20 m bands ≤~4 DN), with MTF-deconvolution OFF so noise and PSF are not re-applied, and (d) remaining limitations are known, dispositioned and risk-tracked.

2. Data package inventory (tailored DRL)#

DRD (ECSS-E-ST-40C Rev.1 / Q-ST-80C)

Document

Status

SSS (Annex B)

sss.md

issued

IRD (Annex C)

ird.md (requirements owned by the SRS)

issued

SRS (Annex D)

srs.md — 51 requirements

issued

ICD (Annex E)

icd.md

issued

SDD (Annex F)

sdd/ incl. traceability

issued

DJF

djf.md — 11 decision records

issued

SRelD/SRN (Annex G)

srn.md + CHANGELOG.md

issued

SUM (Annex H)

sum.md

issued

SVerP + SValP (Annexes I, J)

vv/plan.md (combined)

issued

SUITP (Annex K)

vv/suitp.md

issued

SVR (Annex M)

vv/report.md + vv/s2_l1b_e2e.md

issued

SUITR

vv/suitr.md

issued

SRF (Annex N)

srf.md

issued

SDP (Annex O) incl. tailoring record

sdp.md

issued

SRevP (Annex P)

srevp.md

issued

SPAP (Q-ST-80C)

spa-plan.md

issued

Risk register

risk-register.md

issued

CIDL (M-ST-40C)

cidl.md

issued

SCF

scf.md

issued

ATBD (mission-specific)

atbd/atbd.md — issued v1.0

issued

DPM (mission-specific)

dpm/

issued

SMP (Annex T)

tailored out — folded into the SDP (maintenance = same MR/CI process)

tailored

3. Verification & validation summary#

  • Requirements: 51 baselined; all realized requirements verified PASS by the cited method (V&V report §4). Deferred: REQ-FUNC-043/053/062; cancelled: REQ-FUNC-090 — all with recorded rationale.

  • Unit/integration campaign: 206 collected cases, 201 passed, 0 failed, 5 env-gated skips that pass on the SDE (SUITR); latest main pipeline #31052 green.

  • S2 L1B E2E (reverse-chain) acceptance criteria — all 5 met (S2 L1B E2E validation): primary — the reverse-chain reconstruction agrees with the reference ESA L0 img on the 10/20 m bands (≤~4 DN), with MTF-deconvolution OFF so noise and PSF are not re-applied. Supporting shared-machinery checks: L0plus codec bit-exact 13/13 all bands (decode(L0plus)==L1A round-trip); ISP self-parse 100 % (30 642 packets); PSFD naming round-trip on every product; EOQC OK on both L0 forms.

  • Release evidence: GitLab Release v0.3.0; products + run reports frozen in the generic package registry e2e-s2-l1b/0.3.0; docs site published (strict-build pipeline).

4. Open items & dispositions#

Item

Disposition

Real image-ISP .bin accounting impossible (bucket GET-403)

informative-only by design; recorded verbatim in isp_structural.json (RSK-01)

Codec not interoperable with reference decoders (§4.5.3 divergence)

documented in ICD-IF-C122 + DEC-05; packaged decoder is the consumer path (RSK-02)

Public L1A is DN-scaled

accepted; absolute-radiometry confidence from the Synthetic L0 matching the reference ESA L0 img (≤~4 DN on the 10/20 m bands) (RSK-04)

Deferred requirements (043/053/062)

carried to a future cycle; no impact on the qualified scope

5. Conclusion#

The v0.3.0 baseline is qualified for its stated purpose — the reverse chain: from a Sentinel-2B L1B, reconstruct L1A → L0plus → Synthetic L0 by inverting the operational L0→L1B chain (offset, relative-response/PRNU, dark, un-bin, SWIR re-stage, defective, crosstalk, on-board-eq; MTF-deconvolution OFF), validated against the reference ESA L0 img (10/20 m bands ≤~4 DN): the tailored DRL is complete, the verification campaign is green with the acceptance criteria met on S2 data, and all residual limitations are external-data properties, dispositioned and risk-tracked. No blocking findings.