How harshgounder/lunar-tie-r0 works
lunar-tie-r0 is a classical Chandrayaan-2 lunar image-registration pipeline that produces subpixel tie points and tier-stamped quality results. The principal workflow is to prepare an image pair, detect and match features, estimate and refine their geometry, assess spatial coverage, and issue conformal accept/reject/abstain decisions; pipeline.py is the documented runtime entry point. The README documents these capabilities, while sampled source confirms selected ingest, geometry, and algorithm details. Wiring for unsampled modules is shown only where the README describes the pipeline spine or module role.
Main entry points: pipeline.py Pair pipeline · realdata.py Crop-pair ingest · ch2_ingest.py Product ZIP ingest · geometry.py Geometry lookup
Components 22
Ingest and Georeference
- PDS3 labels
src/lunar_tie/pds3label.py - PDS4 labels
src/lunar_tie/pds4label.py - Product ZIP ingest
src/lunar_tie/ch2_ingest.py - Crop-pair ingest
src/lunar_tie/realdata.py - Geometry lookup
src/lunar_tie/geometry.py - Sun-angle parsing
src/lunar_tie/sun_angles.py - SPICE angles
src/lunar_tie/spice_angles.py
Image Preparation
- Image masks
src/lunar_tie/masking.py - Memory-map tiling
src/lunar_tie/tiling.py - Photometric normalization
src/lunar_tie/photometric.py - Common-GSD resampling
src/lunar_tie/resample.py
Tie Registration
- Pair pipeline
src/lunar_tie/pipeline.py - Feature detect and match
src/lunar_tie/detect.py - Geometric consensus
src/lunar_tie/consensus.py - Subpixel refinement
src/lunar_tie/subpixel.py
Quality and Results
- Coverage and ANMS
src/lunar_tie/coverage.py - Conformal quality gates
src/lunar_tie/conformal.py - Panel and ties
Other components
- Pipeline user
- CH-2 products
- Geometry CSV
- SPICE kernels
Connections 21
- Pipeline user → Pair pipeline: runs pair
- CH-2 products → Product ZIP ingest: provides archive
- Product ZIP ingest → PDS4 labels: parses label
- Crop-pair ingest → Product ZIP ingest: discovers products
- Crop-pair ingest → Geometry lookup: queries geometry
- Geometry CSV → Geometry lookup: supplies grid
- Crop-pair ingest → Pair pipeline: writes pair manifest
- Pair pipeline → PDS3 labels: uses labels
- Pair pipeline → Image masks: runs masking
- Pair pipeline → Memory-map tiling: runs tiling
- Pair pipeline → Photometric normalization: normalizes images
- Pair pipeline → Common-GSD resampling: matches shapes
- Pair pipeline → Feature detect and match: matches features
- Pair pipeline → Geometric consensus: fits geometry
- Pair pipeline → Subpixel refinement: refines ties
- Pair pipeline → Coverage and ANMS: checks spread
- Pair pipeline → Conformal quality gates: gates metrics
- Pair pipeline → Sun-angle parsing: uses sun data
- SPICE kernels → SPICE angles: provides kernels
- Pair pipeline → SPICE angles: extracts angles
- Pair pipeline → Panel and ties: writes results