# Kimera

Visual Inertial Odometry with SLAM capabilities and 3D Mesh generation.

Repository: https://github.com/MIT-SPARK/Kimera-VIO
Canonical: https://ross.abutalabs.com/products/kimera
Language: C++
License: BSD-2-Clause
License Family: permissive
Topics: slam, localization, mapping, euroc-dataset, robotics, visual-inertial-odometry, state-estimation, vio, reconstruction, cpu, real-time
Last push: 2026-08-06T21:51:02+00:00

## Health v2 (maintenance only)
Score: 66/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 96, release rhythm 8, longevity 100
- inputs: {"age_days": 2542, "days_push": 27, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1905, forks 474 (observed 2026-08-28T04:05:51.877262+00:00)

## What it is
Kimera is a modular C++ library for real-time metric-semantic SLAM that builds semantically annotated 3D meshes from camera and inertial data. It comprises modules for visual-inertial odometry, robust pose graph optimization, 3D mesh generation, and semantic 3D reconstruction, and is ROS-enabled while running on CPU.

## Use cases
- run visual-inertial odometry on robot camera and IMU data
- build a semantically annotated 3D mesh of an environment in real time
- perform robust pose graph optimization for SLAM loop closures
- generate per-frame and multi-frame 3D meshes from stereo or RGB-D sensors
- do metric-semantic SLAM research on a CPU without a GPU
- integrate a modular SLAM pipeline into a ROS project

## When to choose
- you need a real-time, CPU-only SLAM or VIO pipeline in C++
- you want semantic 3D reconstruction rather than just sparse point clouds
- your stack is ROS-based and you want modular components you can swap
- you are doing robotics or perception research and need a citable, well-known baseline

## When to avoid
- you need a lightweight SLAM solution without ROS dependencies
- you require GPU-accelerated dense reconstruction at scale
- you need a actively developed project with frequent releases (last release was 2021)
- your use case is 2D mapping or non-visual localization such as LiDAR-only SLAM

## Facets
- artifact type: library
- maturity: maintenance
- function: computer-vision, simulation, graphics, machine-learning
- domain: robotics, computer-vision, autonomous-vehicles, simulation
- platform: cpp
- tags: slam, visual-inertial-odometry, 3d-reconstruction, semantic-mapping, mesh-generation, ros, pose-graph-optimization, linux

## Member repositories
- MIT-SPARK/Kimera-VIO (main) score 66
- MIT-SPARK/Kimera (docs) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:51.877262+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 2026-08-30T02:53:01.253376+00:00, confidence not recorded.
  - readme: https://github.com/MIT-SPARK/Kimera-VIO (fetched 2026-08-28T04:05:51.877262+00:00, sha 963d9137d7df)
- Data as of 2026-08-30T08:39:29.467469+00:00.