Skip to content

Latest commit

 

History

History
79 lines (67 loc) · 10.4 KB

README.md

File metadata and controls

79 lines (67 loc) · 10.4 KB
Raw

Awesome-LiDAR-Camera-Calibration

Awesome

A Collection of LiDAR-Camera-Calibration Papers, Toolboxes and Notes.

Outline

0. Introduction

For applications such as autonomous driving, robotics, navigation systems, and 3-D scene reconstruction, data of the same scene is often captured using both lidar and camera sensors. To accurately interpret the objects in a scene, it is necessary to fuse the lidar and the camera outputs together. Lidar camera calibration estimates a rigid transformation matrix (extrinsics, rotation+translation, 6 DoF) that establishes the correspondences between the points in the 3-D lidar plane and the pixels in the image plane.

Example

1. Target-based methods

Paper Target Feature Optimization Toolbox Note
Extrinsic Calibration of a Camera and Laser Range Finder (improves camera calibration), 2004 checkerboard C:Plane (a), L: pts in plane (m) point-to-plane CamLaserCalibraTool CN
Fast Extrinsic Calibration of a Laser Rangefinder to a Camera, 2005 checkerboard C: Plane (a), L: Plane (m) plane(n/d) correspondence, point-to-plane LCCT *
Extrinsic calibration of a 3D laser scanner and an omnidirectional camera, 2010 checkerboard C: plane (a), L: pts in plane (m) point-to-plane cam_lidar_calib *
LiDAR-Camera Calibration using 3D-3D Point correspondences, 2017 cardboard + ArUco C: 3D corners (a), L: 3D corners (m) ICP lidar_camera_calibration *
Reflectance Intensity Assisted Automatic and Accurate Extrinsic Calibration of 3D LiDAR and Panoramic Camera Using a Printed Chessboard, 2017 checkerboard C: 2D corners (a), L: 3D corners (a) PnP, angle difference ILCC *
Extrinsic Calibration of Lidar and Camera with Polygon, 2018 regular cardboard C: 2D edge, corners (a), L: 3D edge, pts in plane (a) point-to-line, point-inside-polygon ram-lab/plycal *
Automatic Extrinsic Calibration of a Camera and a 3D LiDAR using Line and Plane Correspondences, 2018 checkerboard C: 3D edge, plane(a), L: 3D edge, pts in plane (a) direcion/normal, point-to-line, point-to-plane Matlab LiDAR Toolbox *
Improvements to Target-Based 3D LiDAR to Camera Calibration, 2020 cardboard with ArUco C: 2d corners (a), L: 3D corners (a) PnP, IOU github *
ACSC: Automatic Calibration for Non-repetitive Scanning Solid-State LiDAR and Camera Systems, 2020 checkerboard C: 2D corners (a), L: 3D corners (a) PnP ACSC *
Automatic Extrinsic Calibration Method for LiDAR and Camera Sensor Setups, 2021 cardboard with circle & Aruco C: 3D points (a), L: 3D points (a) ICP velo2cam_ calibration *
Single-Shot is Enough: Panoramic Infrastructure Based Calibration of Multiple Cameras and 3D LiDARs,2021 IROS panoramic infrastructure C: CCTag(a), L: corner points and vectors(a) PnP, ICP multiple-cameras-and-3D-LiDARs-extrinsic-calibration *
Joint Camera Intrinsic and LiDAR-Camera Extrinsic Calibration, 2023, ICRA chessboard + circles C: chessboard corners, circle centers, L: circle centers PnP, intrinsic constraints (corners, cirlce center) OpenCalib/JointCalib CN
LCE-Calib: Automatic LiDAR-Frame/Event Camera Extrinsic Calibration With a Globally Optimal Solution checkboard 3D point-to-3D plan, 3D point-to-3D edge point-to-plane, point-to-edge LCECalib *

C: camera, L: LiDAR, a: automaic, m: manual

2. Targetless methods

2.1. Motion-based methods

Paper Feature Optimization Toolbox Note
LiDAR and Camera Calibration Using Motions Estimated by Sensor Fusion Odometry, 2018 C: motion (ICP), L: motion (VO) hand-eye calibration * *

2.2. Scene-based methods

2.2.1. Traditional methods

Paper Feature Optimization Toolbox Note
Automatic Targetless Extrinsic Calibration of a 3D Lidar and Camera by Maximizing Mutual Information, 2012 C：grayscale, L: reflectivity mutual information, BB steepest gradient ascent Extrinsic Calib *
Automatic Calibration of Lidar and Camera Images using Normalized Mutual Information, 2013 C:grayscale, L: reflectivity, noraml normalized MI, particle swarm * *
Automatic Online Calibration of Cameras and Lasers, 2013 C: Canny edge, L: depth-discontinuous edge correlation, grid search * *
SOIC: Semantic Online Initialization and Calibration for LiDAR and Camera, 2020 semantic centroid PnP * *
A Low-cost and Accurate Lidar-assisted Visual SLAM System, 2021 C: edge(grayscale), L: edge (reflectivity, depth projection) ICP, coordinate descent algorithms CamVox *
Pixel-level Extrinsic Self Calibration of High Resolution LiDAR and Camera in Targetless Environments,2021 C:Canny edge(grayscale), L: depth-continuous edge point-to-line, Gaussian-Newton livox_camera_calib *
CRLF: Automatic Calibration and Refinement based on Line Feature for LiDAR and Camera in Road Scenes, 2021 C:straight line, L: straight line perspective3-lines (P3L) * CN
Fast and Accurate Extrinsic Calibration for Multiple LiDARs and Cameras,2021 C:Canny edge(grayscale), L: depth-continuous edge point-to-plane, point-to-edge mlcc *
General, Single-shot, Target-less, and Automatic LiDAR-Camera Extrinsic Calibration Toolbox,ICRA 2023 C:keypoints, L: keypoints on intensity images (by SuperGlue) reprojection error minimization, RANSAC direct_visual_lidar_calibration *
Calib-Anything: Zero-training LiDAR-Camera Extrinsic Calibration Method Using Segment Anything, 2023 C: Segmentation, L: normal estimation, plane fitting & euclidean cluster, intensity normalization random search, grid search OpenCalib *

2.2.2. Deep-learning methods

Paper Toolbox Note
RegNet: Multimodal sensor registration using deep neural networks, 2017,IV regnet *
CalibNet: Geometrically supervised extrinsic calibration using 3d spatial transformer networks,2018,IROS CalibNet *
LCCNet: Lidar and Camera Self-Calibration using CostVolume Network,2021,CVPRW LCCNet *
Calib-Anything: Zero-training LiDAR-Camera Extrinsic Calibration Method Using Segment Anything,2023 Calib-Anything *

3. Other toolboxes

Toolbox Introduction Note
Apollo sensor calibration tools targetless method, no source code CN
Autoware camera lidar calibrator pick points mannually, PnP *
Autoware calibration camera lidar checkerboard, similar to LCCT CN
livox_camera_lidar_calibration pick points mannually, PnP *
OpenCalib target-based, target-less, mannual OpenCalib: A Multi-sensor Calibration Toolbox for Autonomous Driving
tier4/CalibrationTools target-based, mannual *