mapping-localization
Operate hloc pair generation, COLMAP/pycolmap mapping, triangulation, SfM localization, InLoc localization, and pose outputs.
- License 许可
- Apache-2.0
- 最近更新
- 2026-10-08
- Stars
- ★ 11
- Forks
- 3
01 前言 Front Matter
- name
- mapping-localization
- license
- Apache 2.0
- metadata
- { "disco-role": "operating" }
- description
- Operate hloc pair generation, COLMAP/pycolmap mapping, triangulation, SfM localization, InLoc localization, and pose outputs.
- disable-model-invocation
- true
02 资源文件 4
- 参考文档references/api-and-cli.md9.0 KB
- 参考文档references/data-contracts.md7.5 KB
- 参考文档references/troubleshooting.md7.7 KB
- 参考文档references/workflows.md12.0 KB
03 SKILL.md 正文
mapping-localization
Use this sub-skill when the task is to turn prepared hloc features, retrieval descriptors, matches, image lists, and/or existing camera poses into image pairs, COLMAP/pycolmap sparse models, triangulated reference models, or query poses.
Route here for
- Pair files from exhaustive image sets, retrieval descriptors, covisibility, or known poses.
hloc.reconstruction.main/python -m hloc.reconstructionsparse reconstruction.hloc.triangulation.main/python -m hloc.triangulationtriangulation from known camera poses.hloc.localize_sfm.main/python -m hloc.localize_sfmquery localization against a COLMAP/pycolmap SfM model.hloc.localize_inloc.main/python -m hloc.localize_inlocInLoc-style RGB-D scan localization.- NVM-to-COLMAP import planning with
python -m hloc.colmap_from_nvm.
Route away
- Feature extractor, global descriptor, dense matcher, sparse matcher, or built-in config selection details: use
../feature-retrieval/. - Benchmark-specific directory/download/submission layouts such as Aachen, 4Seasons, RobotCar, or complete InLoc dataset setup: use
../dataset-pipelines/. - Creating custom extractors/matchers or external HDF5 feature/match exports: use
../custom-interop/.
Required operating contract
Before running mapping or localization, identify these artifacts explicitly:
- Image root and any image/query list files.
- Pair or retrieval file, with exactly two whitespace-separated image names per line.
- HDF5 local features and HDF5 matches with hloc-compatible group/dataset names.
- For reconstruction: output SfM directory, image directory, feature file, match file, and pair file.
- For triangulation/localization: a reference COLMAP/pycolmap model directory with binary model files.
- For query localization: query intrinsics list and output pose path.
- For InLoc-style localization: retrieval pairs, query/database feature and match files, and scan/alignment files in the dataset tree.
Run the bundled validator when accepting user-supplied paths or text files:
python sub-skills/mapping-localization/scripts/validate_hloc_inputs.py \
--workflow localize-sfm \
--query-list queries_with_intrinsics.txt \
--pairs retrieval.txt \
--features features.h5 \
--matches matches.h5 \
--reference-model sfm_model
References
04 同仓库更多 Skills
obsidian-clipper-template-creator
Guide for creating templates for the Obsidian Web Clipper. Use when you want to create a new clipping template, understand available variables, or format clipped content.
obsidian-clipper-template-creator
Guide for creating templates for the Obsidian Web Clipper. Use when you want to create a new clipping template, understand available variables, or format clipped content.
python-api
Programmatic Object-Detection-Metrics API guidance for BoundingBox, BoundingBoxes, Evaluator, enums, coordinate conversion, AP/mAP semantics, and source-style import diagnostics.
file-evaluation
Folder-based PASCAL VOC AP/mAP evaluation from ground-truth and detection text files.
object-detection-metrics
Guides legacy Object-Detection-Metrics PASCAL VOC AP/mAP evaluation from detection text folders and direct Python API classes.
nx-workspace-patterns
Configure and optimize Nx monorepo workspaces. Use when setting up Nx, configuring project boundaries, optimizing build caching, or implementing affected commands.