Parallel holographic generation of multiple Hough transform slices

Area of Science:

• Optics and Photonics
• Computer Vision
• Image Processing

Background:

• The Hough transform (HT) is crucial for pattern recognition but computationally intensive.
• Many applications only need specific slices of the 2-D Hough space, not the full computation.

Purpose of the Study:

• To develop a parallel holographic optical architecture for computing select Hough transform (HT) slices.
• To demonstrate the architecture's efficiency and applicability in industrial quality control.

Main Methods:

• Utilized computer-generated holograms and holographic optical elements.
• Developed a novel holographic optical architecture for parallel processing.
• Applied the architecture to detect printing defects like skew on cigarette packs.

Main Results:

• Successfully demonstrated parallel computation of several theta slices of the Hough transform (HT).
• Laboratory results confirmed the feasibility and effectiveness of the holographic approach.
• Showcased potential for generating the entire 2-D Hough space using holographic elements.

Conclusions:

• The proposed holographic optical architecture offers an efficient parallel method for computing HT slices.
• This approach significantly reduces computational load for applications requiring partial Hough space analysis.
• The technology shows promise for real-time industrial inspection and defect detection.