LiDAR Forensics v1.0.0
- Deterministic multistream timing diagnostics for LiDAR recording failures.
- Built-in demonstrations for normal operation, a LiDAR-only stall, and a recording-wide gap.
- Deterministic synthetic public-safe datasets and evidence artifacts.
- Finding register with primary incidents, related findings, and evidence-based classification.
- CSV, JSON, standalone HTML, and Markdown exports.
- Optional AI-assisted engineering brief; the core workflow remains API-key-free.
- 24 passing automated tests.
- No proprietary RAW data is included.
The public application detects, characterizes, and reports failure signatures. It does not perform full point-cloud reconstruction or claim that every damaged recording is recoverable.
During the original forensic investigation, surviving post-failure sensor measurements were used to reconstruct usable point-cloud geometry that standard processing had treated as permanently lost. Geometric recovery depends on which measurements, timestamps, and trajectory information survived the failure.