How head-inthe-cloud/palms-plane-based-accessible-indoor-localization-using-mobile-smartphones works
PALMS and PALMS+ are indoor-localization systems that turn floor plans and either LiDAR/plane observations or RGB-derived depth and point clouds into a location estimate; PALMS+ also supports sequential tracking. The README documents the three entry-point workflows and the major capabilities, while sampled source and the SOURCE INDEX support several internal module relationships. Runner-to-module calls and some unsampled stages are not visible, so those links are omitted rather than inferred.
Main entry points: dataset.py Dataset loaders · alignment.py Scale alignment · simulator.py Tracking simulator · CES.py CES layout matching
Components 20
PALMS localization
- PALMS runner
test_palms.py - PALMS+ single-shot
test_pp.py - PALMS+ sequential
test_pp_seq.py - Tracking simulator
palms_src/simulator.py - Particle filter
palms_src/particle_filter.py - CES layout matching
layout_matching_module/CES.py
Image observations
- Monocular depth
observation_module/depth_pro_estimator.py - Depth processing
observation_module/depth.py - Observation builder
observation_module/observation_module.py - Scale alignment
observation_module/alignment.py - Ground plane
observation_module/ground_plane.py - Point-cloud projection
observation_module/pointcloud.py
Data and mapping
- Dataset loaders
utils/dataset.py - Dataset constants
pp_src/const.py
Shared utilities
- File and map I/O
utils/file_io.py - Camera geometry
utils/camera_utils.py - Geometry operations
utils/geometry_utils.py - Image sampling
utils/image_sampling.py - Visualization
utils/visualization.py
Other components
- Researcher
Connections 30
- Researcher → PALMS runner: runs
README.md - Researcher → PALMS+ single-shot: runs
README.md - Researcher → PALMS+ sequential: runs
README.md - Tracking simulator → Particle filter: invokes
palms_src/simulator.py - Observation builder → Scale alignment: uses
observation_module/observation_module.py - Observation builder → Depth processing: uses
observation_module/observation_module.py - Monocular depth → Depth processing: uses
observation_module/depth_pro_estimator.py - Monocular depth → File and map I/O: uses
observation_module/depth_pro_estimator.py - Monocular depth → Camera geometry: uses
observation_module/depth_pro_estimator.py - Scale alignment → Ground plane: uses
observation_module/alignment.py - Scale alignment → Point-cloud projection: uses
observation_module/alignment.py - Scale alignment → Image sampling: uses
observation_module/alignment.py - Scale alignment → Camera geometry: uses
observation_module/alignment.py - Point-cloud projection → Geometry operations: uses
observation_module/pointcloud.py - Dataset loaders → Depth processing: masks depths
utils/dataset.py - Dataset loaders → File and map I/O: uses
utils/dataset.py - Dataset loaders → Camera geometry: uses
utils/dataset.py - Dataset loaders → Image sampling: uses
utils/dataset.py - Dataset loaders → Dataset constants: uses
utils/dataset.py - Tracking simulator → Geometry operations: uses
palms_src/simulator.py - Tracking simulator → Visualization: uses
palms_src/simulator.py - Particle filter → Geometry operations: uses
palms_src/particle_filter.py - CES layout matching → File and map I/O: uses
layout_matching_module/CES.py - CES layout matching → Geometry operations: uses
layout_matching_module/CES.py - CES layout matching → Visualization: uses
layout_matching_module/CES.py - PALMS runner → Tracking simulator: runs tracking
- PALMS+ single-shot → Observation builder: builds observations
- PALMS+ sequential → Observation builder: builds observations
- PALMS+ single-shot → CES layout matching: matches layout
- PALMS+ sequential → Tracking simulator: tracks posterior