Camera Calibration using Zhang's Method

A Python implementation of Zhang's camera calibration technique using a planar checkerboard pattern. This project provides tools for estimating both intrinsic and extrinsic camera parameters through checkerboard corner detection, homography computation, and non-linear optimization.

Features

• Automatic checkerboard corner detection and sub-pixel refinement
• Intrinsic camera parameter estimation (focal length, principal point)
• Extrinsic parameter computation (rotation, translation) for each calibration image
• Lens distortion correction (radial distortion coefficients k₁, k₂)
• Visualization tools for calibration quality assessment
• Image undistortion capabilities

Overview

The calibration pipeline follows these key steps:

1. Image Acquisition and Corner Detection

   • Process multiple JPEG calibration images
   • Automatic checkerboard corner detection using OpenCV
   • Sub-pixel corner refinement for improved accuracy

2. Initial Parameter Estimation

   • Homography computation using Direct Linear Transform (DLT)
   • Intrinsic parameter extraction from homography constraints
   • Per-image extrinsic parameter estimation (R, t)

3. Non-Linear Parameter Refinement

   • Joint optimization of all parameters
   • Minimization of reprojection error
   • Estimation of lens distortion coefficients
   • Bundle adjustment for improved accuracy

4. Visualization and Validation

   • Image undistortion using calibrated parameters
   • Overlay of detected vs. reprojected corners
   • Reprojection error visualization and statistics

Installation

Prerequisites

• Python 3.x
• OpenCV
• NumPy
• SciPy

Setup

# Clone the repository
git clone https://github.com/yourusername/camera-calibration.git
cd camera-calibration

# Install required packages
pip install opencv-python numpy scipy

Usage

# Example code snippet for running calibration
from calibration import CameraCalibration

# Initialize calibrator
calibrator = CameraCalibration()

# Add calibration images
calibrator.add_images('path/to/calibration/images/*.jpg')

# Run calibration
calibrator.calibrate()

# Get calibration results
K, dist_coeffs = calibrator.get_parameters()

# Undistort an image
undistorted = calibrator.undistort_image(image)

Results

Original Checkerboard Images

Checkerboard 1	Checkerboard 2

Corner Detection and Reprojection

Reprojection 1	Reprojection 2

Undistorted Results

Undistorted Image 1	Undistorted Image 2

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Citation

If you use this implementation in your research, please cite:

@article{zhang2000flexible,
  title={A flexible new technique for camera calibration},
  author={Zhang, Zhengyou},
  journal={IEEE Transactions on pattern analysis and machine intelligence},
  volume={22},
  number={11},
  pages={1330--1334},
  year={2000},
  publisher={IEEE}
}