# koide3/fast_gicp

A collection of GICP-based fast point cloud registration algorithms

Repository: https://github.com/koide3/fast_gicp
Canonical: https://ross.abutalabs.com/products/fast_gicp
Language: C++
License: BSD-3-Clause
License Family: permissive
Topics: point-cloud, pcl, gicp, icp, cuda, gpu, multithreading, registration, vgicp, scan-matching, python, cpp
Last push: 2025-04-24T01:46:48+00:00

## Health v2 (maintenance only)
Score: 40/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 18, release rhythm 35, longevity 100
- inputs: {"age_days": 2401, "days_push": 497, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1686, forks 372 (observed 2026-08-28T04:05:22.167792+00:00)

## What it is
A C++ library of fast GICP-based point cloud registration algorithms, including multi-threaded GICP, voxelized GICP (VGICP), and CUDA-accelerated VGICP and NDT implementations. It provides PCL-compatible registration interfaces plus Python bindings, allowing drop-in replacement for PCL's GICP.

## Use cases
- align two 3d point clouds fast
- scan matching for lidar slam
- replace pcl gicp with a faster implementation
- gpu-accelerated point cloud registration
- register point clouds from python
- ndt scan matching for odometry
- voxelized gicp for large point clouds

## When to choose
- you need high-throughput point cloud registration with PCL-compatible interfaces
- you want CUDA-accelerated VGICP or NDT for real-time robotics
- you need Python bindings for point cloud alignment
- you are building SLAM or odometry pipelines on Linux

## When to avoid
- you want the fastest, most lightweight option - the author recommends small_gicp instead
- you need Windows support or minimal dependencies
- you only need simple ICP without covariance-based registration

## Facets
- artifact type: library
- maturity: maintenance
- function: computer-vision, image-processing, math, sdk
- domain: robotics, computer-vision, autonomous-vehicles
- platform: cpp, python
- tags: point-cloud-registration, gicp, icp, scan-matching, cuda, pcl, ndt, 3d-slam, lidar, algorithms, linux, macos, gpu, ros

## Member repositories
- koide3/fast_gicp (main) score 40

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:22.167792+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:38:59.304741+00:00, confidence not recorded.
  - readme: https://github.com/koide3/fast_gicp (fetched 2026-08-28T04:05:22.167792+00:00, sha a3addad7d63b)
- Data as of 2026-08-30T08:39:29.467469+00:00.