Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Fisheye-Calib-Adapter: An Easy Tool for Fisheye Camera Model Conversion

Sangjun Lee

TL;DR

The paper addresses the challenge of calibrating across diverse fisheye camera models by introducing the Fisheye-Calib-Adapter (FCA), a tool that directly converts intrinsic parameters between multiple models without needing image data. The method unprojects rays from the input model, then optimizes the output model via nonlinear least squares to align projections, formalized as $i_{out}^* = FCA(i_{in})$. It covers six widely used models (UCM, EUCM, DS, KB, OCC, RT) plus a Custom option, detailing projection/unprojection, initialization, and Jacobians for each, and demonstrates strong accuracy and speed (often under $4$ ms per sample) on Kalibr and OCamCalib datasets, with practical SLAM validation. The results show FCA can serve as a practical bridge across models, improving cross-model compatibility for researchers and enabling recalibration-free parameter transfer, while remaining open-source for broad adoption.

Abstract

The increasing necessity for fisheye cameras in fields such as robotics and autonomous driving has led to the proposal of various fisheye camera models. While the evolution of camera models has facilitated the development of diverse systems in the field, the lack of adaptation between different fisheye camera models means that recalibration is always necessary, which is cumbersome. This paper introduces a conversion tool for various previously proposed fisheye camera models. It is user-friendly, simple, yet extremely fast and accurate, offering conversion capabilities for a broader range of models compared to existing tools. We have verified that models converted using our system perform correctly in applications such as SLAM. By utilizing our system, researchers can obtain output parameters directly from input parameters without the need for an image set and any recalibration processes, thus serving as a bridge across different fisheye camera models in various research fields. We provide our system as an open source tool available at: https://github.com/eowjd0512/fisheye-calib-adapter

Fisheye-Calib-Adapter: An Easy Tool for Fisheye Camera Model Conversion

TL;DR

The paper addresses the challenge of calibrating across diverse fisheye camera models by introducing the Fisheye-Calib-Adapter (FCA), a tool that directly converts intrinsic parameters between multiple models without needing image data. The method unprojects rays from the input model, then optimizes the output model via nonlinear least squares to align projections, formalized as . It covers six widely used models (UCM, EUCM, DS, KB, OCC, RT) plus a Custom option, detailing projection/unprojection, initialization, and Jacobians for each, and demonstrates strong accuracy and speed (often under ms per sample) on Kalibr and OCamCalib datasets, with practical SLAM validation. The results show FCA can serve as a practical bridge across models, improving cross-model compatibility for researchers and enabling recalibration-free parameter transfer, while remaining open-source for broad adoption.

Abstract

The increasing necessity for fisheye cameras in fields such as robotics and autonomous driving has led to the proposal of various fisheye camera models. While the evolution of camera models has facilitated the development of diverse systems in the field, the lack of adaptation between different fisheye camera models means that recalibration is always necessary, which is cumbersome. This paper introduces a conversion tool for various previously proposed fisheye camera models. It is user-friendly, simple, yet extremely fast and accurate, offering conversion capabilities for a broader range of models compared to existing tools. We have verified that models converted using our system perform correctly in applications such as SLAM. By utilizing our system, researchers can obtain output parameters directly from input parameters without the need for an image set and any recalibration processes, thus serving as a bridge across different fisheye camera models in various research fields. We provide our system as an open source tool available at: https://github.com/eowjd0512/fisheye-calib-adapter
Paper Structure (17 sections, 41 equations, 4 figures, 4 tables)

This paper contains 17 sections, 41 equations, 4 figures, 4 tables.

Table of Contents

  1. INTRODUCTION
  2. Related Work
  3. Fisheye Camera Models
  4. Fisheye Camera Model Conversion
  5. Method
  6. Unified Camera Model
  7. Enhanced Unified Camera Model
  8. Double Sphere
  9. Kannala-Brandt Camera Model
  10. OCamCalib Camera Model
  11. Radial-Tangential Distortion Model
  12. Custom Camera Model
  13. Experiment
  14. Evaluation
  15. Application
  16. ...and 2 more sections

Figures (4)

  • Figure 1: System overview
  • Figure 2: For the comparison of parameter error and execution time per sample point $N$ , the results are presented in two parts: (a) parameter error per sample point and (b) execution time per sample point. The notation A-B indicates the conversion from input model A to output model B. The parameter error begins to saturate at approximately $N=30$, and the execution time starts to significantly increase from around $N=1000$.
  • Figure 3: An example of recovered images. The first row is from the Kalibr dataset, and the second row is from the OCamCalib dataset. The input model is fixed as KB, and the output model corresponds to each column of the image.
  • Figure 4: An example of fisheye ORB SLAM. (a) Results obtained using directly acquired EUCM parameters. (b) Results obtained using EUCM parameters acquired through KB-EUCM conversion performed by our model.