Directed Research · ERAU
MTF / Optical
Bench Study
Slanted-edge spatial frequency response analysis comparing monochrome and Bayer color camera sensors across multiple aperture settings. Custom MATLAB scripts measured edge spread function, line spread function, and modulation transfer function with Nyquist-aware interpretation of the results.
Overview
Research Question
Monochrome vs. Bayer — how does color interpolation affect resolving power?
Bayer color cameras demosaic raw sensor data through color filter interpolation, which can soften fine spatial detail. This study measured and compared the actual resolving power of both sensor types across aperture settings to quantify that tradeoff.
Methodology
Slanted-Edge SFR
The ISO 12233 slanted-edge method provides a statistically robust way to compute a camera system's spatial frequency response from a single high-contrast edge target. Oversampling across the slant yields a sub-pixel-accurate edge spread function.
Analysis Pipeline
Step 01
ESF Extraction
Edge Spread Function extracted from the slanted-edge target image. Custom MATLAB script locates the edge, resamples across the slope, and bins pixel values to produce a smooth ESF curve.
Step 02
LSF Derivation
Line Spread Function computed as the derivative of the ESF. A Gaussian smoothing pass suppresses derivative noise before the LSF is passed to the frequency domain.
Step 03
MTF + Nyquist Interpretation
Modulation Transfer Function computed via FFT of the LSF. Results were interpreted relative to the Nyquist frequency of each sensor to give a normalized resolving-power comparison.
Test Variables
Camera Types
Monochrome vs. Bayer CFA
A monochrome sensor captures luminance at every pixel. A Bayer sensor covers each pixel with a red, green, or blue filter and reconstructs full color via demosaicing — at the cost of spatial high-frequency information.
Aperture Settings
Multiple f-stops
Tests were run across multiple aperture settings to capture the interaction between diffraction-limited sharpness at small apertures and aberration-driven softness at wide apertures — isolating sensor-type effects from lens effects.