Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/ethz-asl/kalibr/llms.txt

Use this file to discover all available pages before exploring further.

Kalibr is a robust calibration toolbox developed at ETH Zurich’s Autonomous Systems Lab. It solves challenging calibration problems for robotics and computer vision systems — from intrinsic and extrinsic multi-camera calibration to spatial and temporal camera-IMU calibration — using continuous-time batch estimation over ROS bag recordings.

Installation

Install Kalibr via Docker or build from source in a ROS catkin workspace.

Quickstart

Run your first camera calibration in minutes with an AprilGrid target.

Calibration Guides

Step-by-step guides for camera, camera-IMU, and rolling shutter calibration.

CLI Reference

Full reference for all Kalibr command-line tools and their flags.

​
What Kalibr calibrates

Kalibr handles four distinct calibration problems, each with a dedicated CLI tool:

Multi-Camera Calibration

Intrinsic and extrinsic calibration of camera systems with non-overlapping fields of view.

Camera-IMU Calibration

Spatial and temporal calibration of an IMU with respect to a camera chain.

Multi-IMU Calibration

Calibrate multiple IMUs relative to a base inertial sensor with an aiding camera.

Rolling Shutter Calibration

Full intrinsic calibration including shutter parameters for rolling shutter cameras.

​
Getting started

1

Install Kalibr

Pull the Docker image for your Ubuntu version, or build from source inside a ROS catkin workspace. See the installation guide for details.
2

Print a calibration target

Generate an AprilGrid or checkerboard PDF with kalibr_create_target_pdf and print it at the correct physical size.
3

Record a ROS bag

Move the calibration target in front of your sensors while recording camera and IMU topics to a ROS bag file.
4

Run calibration

Execute the appropriate Kalibr command with your bag file, target configuration, and camera model arguments to produce YAML calibration results and a PDF report.

​
Supported camera models

Kalibr supports a wide range of projection and distortion models out of the box:
Model stringProjectionDistortion
pinhole-radtanPinholeRadial-tangential
pinhole-equiPinholeEquidistant (fisheye)
pinhole-fovPinholeField-of-view
omni-noneOmnidirectionalNone
omni-radtanOmnidirectionalRadial-tangential
eucm-noneExtended UnifiedNone
ds-noneDouble SphereNone
Not sure which model to use? Start with pinhole-radtan for standard cameras and pinhole-equi for wide-angle or fisheye lenses. See Camera Models for guidance.

Build docs developers (and LLMs) love

Get started for freeTalk to us