Evaluation of image matching techniques for optical fiber specklegram sensor analysis

Area of Science:

• Optics and Photonics
• Materials Science
• Sensor Technology

Background:

• Fiber optic sensors are crucial for measuring physical and chemical stimuli.
• Specklegram analysis offers a method to quantify fiber optic sensor responses.
• Microbending transducers modulate fiber status, altering the output speckle field.

Purpose of the Study:

• To quantitatively evaluate various image matching techniques for fiber specklegram sensor analysis.
• To determine the sensitivity and accuracy of different methods in correlating specklegram variations to fiber displacements.
• To identify optimal techniques for detecting subtle changes in fiber status.

Main Methods:

• Acquisition and preprocessing of fiber specklegrams.
• Application of seven distinct image matching approaches, including average intensity, correlation, sum of differences, phase-only correlation, and mutual information.
• Quantitative assessment of sensitivity, linearity, and hysteresis errors for each method.

Main Results:

• Average intensity analysis was found to be unreliable.
• Correlation and sum of differences methods showed sensitivities of ~11 mm⁻¹ and ~14 mm⁻¹, respectively, within a ~0.06 mm range.
• Phase-only correlation and mutual information demonstrated high sensitivities (22 mm⁻¹ and 120 mm⁻¹) for displacements <0.02 mm, with low linearity and hysteresis errors.

Conclusions:

• Image matching techniques, particularly phase-only correlation and mutual information, are effective for analyzing fiber specklegram sensors.
• These advanced methods enable the detection of minute variations in fiber status caused by physical or chemical stimuli.
• The study highlights the potential for enhanced sensitivity and accuracy in fiber optic sensing through optimized image processing.