A contrast-based autofocus algorithm employing various techniques pertinent to MTE 204 - Numerical Methods

by Yun-Ha Jung, Ruoyu Jessen Liang and Russell Wong

This library of code contains all Python scripts that were developed for this project, including scripts that were tested during the development stage but not used for the final implementation of the autofocus algorithm. For more information regarding the important functions for the finalized algorithm, see Appendix B in the report, "Development of a Contrast-Based Autofocus Algorithm using Numerical Methods"

The main function is contained within golden_section.py. The algorithmic steps are as follows:

1. Load image data and focus distance data based on a subject number and calculate sharpness for each image (sharpness_calc.py)
2. Set up a system of equations for determining a set of cubic functions to approximate the sharpness curve as a function of focus distance (cubic_spline.py)
3. Solve this system of equations using Gaussian Elimination (gaussian_elimination.py), thereby calculating the required coefficients for cubic spline interpolation
4. Approximate the maximum of this cubic spline function using the Golden Section Method, which will identify a maximum sharpness value corresponding to an optimal focus distance (golden_section.py)

Algorithmic details can be found in the report, "Development of a Contrast-Based Autofocus Algorithm using Numerical Methods"