Roadmap

v0.1.0 — Initial Release (current)

Goal: Working CLI and Python API for single-pair boresight calibration.

• Planar-neighbourhood feature extraction (k-NN PCA + Weinmann planarity filter)
• Point-to-plane ICP with Levenberg–Marquardt optimisation (6-DOF: Δω, Δφ, Δκ, Δx, Δy, Δz)
• LAS/LAZ I/O via laspy (preserving all non-XYZ attributes)
• RMSE/MAE metrics with coverage fraction reporting
• CLI: lidar-strip-adjust reference.las target.las -o corrected.las
• Unit + integration tests, CI via GitHub Actions
• MIT licence, CITATION.cff, mkdocs documentation

v0.2.0 — Multi-strip Network Adjustment

Goal: Simultaneous least-squares adjustment across N ≥ 3 flight strips.

• Strip graph construction: pair all overlapping strip combinations automatically
• Weighted global normal-equations solver (scipy sparse + LSQR)
• Per-strip residual report and blunder detection (3σ rejection)
• Tie-point / control-point injection (optional ground-truth LAS)
• Parallelised pairwise feature extraction (concurrent.futures)
• --network CLI subcommand for batch input
• Benchmark: ≥10 strips, ≥5M points each, wall time < 60 s on 8 cores

v1.0.0 — Production-Ready SDK

Goal: API stability, cloud-native support, and ecosystem integrations.

• Stable public API with semantic-versioning guarantees
• PDAL pipeline integration: lidar_strip_adjust filter stage (JSON pipeline)
• Cloud I/O: direct read/write from S3, Azure Blob, GCS via fsspec
• LAZ compression support via lazrs/laszip backend
• Per-point confidence scores written back to LAS extra bytes
• Full type stubs (py.typed, mypy strict)
• Sphinx/ReadTheDocs hosted documentation
• Conda-forge package