High-accuracy calibration of low-cost camera using image disturbance factor

Area of Science:

• Computer Vision
• Robotics
• Image Processing

Background:

• Nonindustrial low-cost cameras offer economic advantages but suffer from low resolution and high image noise.
• Existing camera calibration methods fail to achieve high accuracy with these low-cost cameras.

Purpose of the Study:

• To develop a high-accuracy camera calibration method specifically for nonindustrial low-cost cameras.
• To overcome the limitations of conventional calibration techniques when applied to inexpensive imaging devices.

Main Methods:

• Utilized multiscale image analysis to determine initial values and uncertainties of image feature points.
• Incorporated an image disturbance factor for each feature point and established image projection error as a minimum objective function.
• Employed nonlinear optimization to obtain optimal feature point coordinates and subsequently applied Zhang's method for parameter calibration.

Main Results:

• The proposed method demonstrated a significant improvement in calibration accuracy compared to traditional approaches.
• Experimental results indicated that the new method achieves at least three times the calibration accuracy of Zhang's method.
• Validated through both simulative and real-world experiments.

Conclusions:

• The developed high-accuracy planar target-based calibration method effectively addresses the limitations of low-cost cameras.
• This technique offers a viable solution for applications requiring precise camera parameter estimation with affordable imaging hardware.
• The enhanced accuracy opens possibilities for wider adoption of low-cost cameras in metrology and machine vision.