qingchenbenpao/multi-camera-calibration

Photogrammetric Multi-Camera Calibration Toolbox V1

1. Introduction

We implement a generalization of Zhang's widely used single camera calibration algorithm to N cameras. Solutions are found through the overall optimization of the cameras' parameters. The user can choose which subset of cameras' parameters to re-optimize: re-optimize all parameters, fixed intrinsic parameters or fixed lens distortions.

Note: This version of the toolbox can only handle camera configurations with significantly overlapping fields of view. All cameras must be able to observe the entire calibration target simultaneously.

2.1 Calibration Object

The toolbox expects the calibration object to be a standard planar checkerboard pattern. The checkerboard pattern should be asymmetric and of even and odd dimensions. Example: 28 X 17 corners.

2.2 Calibration Image Capture

Capture as many images as necessary to satisfy the following conditions:

• Each camera sees at least 3 fully visible pattern positions.
• Account for as much of each camera’s image frames as possible.
• In 10 to 20 pattern positions the checkerboard is fully visible by all cameras.
• In each image, the checkerboard pattern must be at different orientations relative to the camera.
• The checkerboard should preferably fill at least 20% of the image frame.
• The checkerboard must be at an angle of less than 45 degrees relative to the camera plane.

2.3 Image Naming Convension

PositionXXX_CameraXX.

2. Dependencies

The implementation depends on the image processing, camera calibration and optimization toolboxes of Matlab.

3. How to Cite?

Cite this page.

3. How to Use?

Run script main.m.

Contact

Laura Ribeiro

Tampere University, Finland

[email protected]