limap.cli package

Command line entry points, each run with python -m limap.cli.<name>. A CLI module parses its arguments together with the YAML configuration under cfgs/ (see limap.util.config.load_config()), converts them into the Options dataclass of the corresponding runner and calls it. They are thin by design: the pipelines themselves live in limap.runners package.

Triangulation from a COLMAP model

Complete pipelines that run the frontend and the triangulation in one go, on a scene whose camera poses are already known.

Points and lines

Runner for point-line triangulation from COLMAP.

This is similar to automatic_structure_triangulation.py but with groups disabled. It only triangulates points and lines, skipping group (plane) detection and triangulation.

Usage:

python -m limap.cli.automatic_point_line_triangulation \
    -m <path_to_colmap_model> \
    -i <path_to_images> \
    -o <output_directory> \
    -c cfgs/structure_triangulation/default.yaml

limap.cli.automatic_point_line_triangulation.main()

limap.cli.automatic_point_line_triangulation.run_point_line_triangulation_from_colmap(cfg, model_path, image_path)

Points, lines, groups and wireframe

Runner for structure triangulation from COLMAP.

Triangulates 3D points, lines, and groups (planes) from a COLMAP model. Runs the full pipeline: frontend (detection + matching), point triangulation, line/group triangulation, and optional bundle adjustment.

Usage:

python -m limap.cli.automatic_structure_triangulation \
    -m <path_to_colmap_model> \
    -i <path_to_images> \
    -o <output_directory> \
    -c cfgs/structure_triangulation/default.yaml

limap.cli.automatic_structure_triangulation.main()

limap.cli.automatic_structure_triangulation.run_structure_triangulation_from_colmap(cfg, model_path, image_path)

Frontend

Detection, description and matching alone, writing the two databases that the triangulation modules below consume.

Runner for structure frontend (detection + matching).

This runner takes a COLMAP model and images, runs line/point detection and matching, and outputs:

database.db: COLMAP database with point features and matches
structure_database.db: Structure database with line features and matches

Usage:

python -m limap.cli.structure_frontend \
    --image_dir <path_to_images> \
    --model_path <path_to_colmap_model> \
    --output_dir <output_directory>

The output can then be used with global_line_triangulation.py or incremental_line_triangulation.py.

limap.cli.structure_frontend.main()

Triangulation from a structure database

The second half of the automatic pipelines, taking the frontend output instead of computing it.

Global line triangulation

Runner for global line triangulation.

This runner runs global line triangulation using a pre-computed structure database and COLMAP model.

Usage:

python -m limap.cli.global_line_triangulation \
    --structure_db_path <path_to_structure_database.db> \
    --model_path <path_to_colmap_model> \
    --output_dir <output_directory>

limap.cli.global_line_triangulation.main()

limap.cli.global_line_triangulation.run_global_line_triangulation(structure_db_path: Path, model_path: Path, output_dir: Path, cfg: dict)

Run global line triangulation using pre-computed databases.

Parameters:

structure_db_path – Path to structure_database.db
model_path – Path to COLMAP model directory
output_dir – Output directory for results
cfg – Configuration dictionary with triangulation options

Global structure triangulation

Runner for global structure triangulation.

This runner runs global structure triangulation (lines + groups) using a pre-computed structure database and COLMAP model.

Usage:

python -m limap.cli.global_structure_triangulation \
    --structure_db_path <path_to_structure_database.db> \
    --model_path <path_to_colmap_model> \
    --output_dir <output_directory>

limap.cli.global_structure_triangulation.main()

limap.cli.global_structure_triangulation.run_global_structure_triangulation(structure_db_path: Path, model_path: Path, output_dir: Path, cfg: dict)

Run global structure triangulation using pre-computed databases.

Parameters:

structure_db_path – Path to structure_database.db
model_path – Path to COLMAP model directory
output_dir – Output directory for results
cfg – Configuration dictionary with triangulation options

Incremental line triangulation

Runner for incremental line triangulation.

This runner simulates incremental reconstruction by processing images one at a time, triangulating lines after each “registration”. It requires a pre-computed structure database (from a previous structure_triangulation run).

Usage:

python -m limap.cli.incremental_line_triangulation \
    --structure_db_path <path_to_structure_database.db> \
    --model_path <path_to_colmap_model> \
    --output_dir <output_directory>

limap.cli.incremental_line_triangulation.main()

limap.cli.incremental_line_triangulation.run_incremental_line_triangulation(structure_db_path: Path, model_path: Path, output_dir: Path, cfg: dict)

Run incremental line triangulation using pre-computed databases.

This simulates incremental SfM by: 1. Loading the structure database (with detections and matches) 2. Processing images one at a time in order 3. Triangulating lines incrementally as each image is “registered”

Parameters:

structure_db_path – Path to structure_database.db
model_path – Path to COLMAP model directory
output_dir – Output directory for results
cfg – Configuration dictionary with triangulation options

Incremental structure triangulation

Runner for incremental structure triangulation.

This runner runs incremental structure triangulation (points, lines, groups) using pre-computed databases. It requires a pre-computed structure database and a COLMAP database with point correspondences.

Usage:

python -m limap.cli.incremental_structure_triangulation \
    --structure_db_path <path_to_structure_database.db> \
    --db_path <path_to_colmap_database.db> \
    --model_path <path_to_colmap_model> \
    --output_dir <output_directory>

limap.cli.incremental_structure_triangulation.main()

limap.cli.incremental_structure_triangulation.run_incremental_structure_triangulation(structure_db_path: Path, db_path: Path, model_path: Path, output_dir: Path, cfg: dict)

Run incremental structure triangulation using pre-computed databases.

Parameters:

structure_db_path – Path to structure_database.db
db_path – Path to COLMAP database.db (for point correspondences)
model_path – Path to COLMAP model directory
output_dir – Output directory for results
cfg – Configuration dictionary with triangulation options

Incremental SfM

Run structure-aware incremental SfM.

This pipeline performs full incremental SfM with hybrid point+line registration, structure triangulation (points, lines, groups), and structure BA.

Usage:

python -m limap.cli.structure_incremental_sfm \
    --db_path <path_to_colmap_database.db> \
    --structure_db_path <path_to_structure_database.db> \
    --image_path <path_to_images> \
    --output_dir <output_directory>

limap.cli.structure_incremental_sfm.main()

limap.cli.structure_incremental_sfm.run_structure_incremental_sfm(db_path: Path, structure_db_path: Path, image_path: Path, output_dir: Path, cfg: dict)

Run structure-aware incremental SfM.

Parameters:

db_path – Path to COLMAP database.db
structure_db_path – Path to structure_database.db
image_path – Path to image directory
output_dir – Output directory for results
cfg – Configuration dictionary