OVERVIEW
You have recently learned about the convolution sum that serves as the basis of the FIR filter difference equation. The filter coefficient sequence {????} - equivalent to the filter's impulse response h[??] - may be viewed as a one-dimensional moving window that slides over the input signal ??[??] to compute the output signal ??[??] at each time step. Extending the moving window concept to a 2-D array that slides over an image pixel array provides a useful and popular way to filter an image.
In this lab project you will implement two types of moving-window image filters, one based on convolution and the other based on the median value of the pixel grayscale values spanned by the window. You will also gain experience with the built-in image convolution filter imfilter.
OUTLINE
1. Develop and test a 3x3 median filter
2. Develop and test a 3x3 convolution filter
3. Evaluate the median and convolution filters to reduce noise while preserving edges
4. Study the behavior of various 3x3 convolution filter kernels for smoothing, edge detection, and sharpening
5. Learn how to use imfilter to convolution-filter color images, and study the various mechanisms offered by imfilter to deal with boundary effects