Reflection holographic filters for compacting optical processors.

Area of Science:

• Optics and Photonics
• Image Processing
• Holography

Background:

• Optical character recognition (OCR) systems often rely on matched filters for pattern matching.
• Holographic filters, particularly those using transmission holograms, have been explored for their potential in OCR.
• Developing filters that offer high discrimination and positional tolerance is crucial for robust character recognition.

Purpose of the Study:

• To investigate the efficacy of reflection hologram filters for character recognition tasks.
• To compare the performance of reflection filters with existing transmission hologram filters.
• To assess the advantages of using reflection filters in terms of processor size and cost.

Main Methods:

• Character recognition experiments were conducted using reflection matched filters.
• Reflection filters were fabricated using volume holograms recorded on photographic emulsion.
• The filters were processed and integrated into a one-lens optical processor.

Main Results:

• Reflection filters demonstrated good discrimination capabilities among similar characters.
• These filters exhibited significant tolerance to variations in character positioning.
• Performance was found to be comparable to that of transmission hologram filters.
• The use of reflection filters allowed for a one-lens processor, halving the size and cost.

Conclusions:

• Reflection hologram filters are a viable and effective alternative for character recognition.
• They offer comparable or superior performance to transmission filters with added benefits.
• The one-lens processor design significantly reduces system complexity, size, and cost.