Global Calibration of Multiple Cameras Based on Sphere Targets

Area of Science:

• Computer Vision
• Metrology
• Optical Engineering

Background:

• Accurate vision measurement relies heavily on precise global calibration of multi-camera systems.
• Existing methods can be complex and may not be suitable for all industrial or research applications.

Purpose of the Study:

• To develop a novel, accurate, and flexible global calibration method for multi-camera systems.
• To improve the reconstruction accuracy of sphere centers using a parameter equation for sphere projection.

Main Methods:

• Utilized groups of sphere targets observed by individual cameras and a precision auxiliary camera.
• Employed a parameter equation to model sphere projection for accurate sphere center reconstruction.
• Conducted theoretical analysis and computer simulations to validate the method's accuracy and influencing factors.

Main Results:

• Simulation results demonstrated significant improvement in reconstruction accuracy using the parameter equation.
• Experimental validation on a two-camera system achieved a root mean squared error within 0.14 mm for a 578 mm measurement.
• The method proved to be simple to operate and flexible, particularly for onsite calibrations without a common field of view.

Conclusions:

• The proposed global calibration method offers high accuracy and robustness for multi-camera systems.
• The parameter equation for sphere projection is key to enhancing reconstruction precision.
• This technique is well-suited for practical applications requiring flexible and accurate vision measurements.