Homography estimation by two PClines Hough transforms and a square-radial checkerboard pattern

Area of Science:

• Optical metrology
• Computer vision
• Geometric transformations

Background:

• Homography matrix is essential for 3D surface imaging in structured-light projection.
• Accurate homography estimation is critical for metrology applications.
• Existing methods may lack straightforwardness or efficiency.

Purpose of the Study:

• To propose a straightforward homography estimation method.
• To introduce a novel square-radial checkerboard pattern for generating concurrent lines.
• To validate the method experimentally for optical metrology.

Main Methods:

• Utilizing a novel square-radial checkerboard pattern to generate three sets of concurrent lines.
• Employing two PClines Hough transforms for detecting lines and points of interest.
• Explaining concepts like parallel coordinates and Hough transforms.

Main Results:

• Successful detection of lines and points of interest using the proposed pattern and Hough transforms.
• Demonstration of a straightforward and effective homography estimation process.
• Experimental verification of the method's usefulness in optical metrology.

Conclusions:

• The proposed method offers a simple and effective approach to homography estimation.
• The novel square-radial checkerboard pattern facilitates accurate line and point detection.
• This technique is valuable for 3D surface imaging and other optical metrology tasks.