EE424 Image Processing Lab

Purpose: Filtering and edge detection in digital images.

Introduction

In this lab we filter a new type of digital signal, two dimensional digital images. The lab shows how to display black and white intensity images. The lab introduces the concepts of high pass and low pass filtering of images and how this filtering affects the frequency content and visual quality of an image. The problems of noise removal and edge enhancement are addressed.

Validation sheets for  report.

Overview

An image can be represented as a function f(x,y) of two continuous variables representing the horizontal and vertical coordinates of a point in space. For monochrome images, f(x,y) is a scalar function of the intensity of the two spatial variables. For color images, f(x,y) is a vector function that denotes the color of the point (e.g. it could have three values for red, green and blue in the RGB coordinate system). Monochrome images are displayed using black and white and shades of gray, thus they are called "grayscale" images.

A sampled grayscale image can be represented as a two-dimensional array of number of the form:

The values of M and N are usually a power of 2 such as 256 or 512 since this makes many image processing operations such as those involving FFTs much faster. Matlab represents an image as a matrix with M rows and N columns. The matrix entry at (m,n) is the sample value f(m,n) called a pixel (picture element).

Grayscale images show the intensity of each pixel, which must always be finite and non-negative:

Grayscale images commonly use eight or sixteen bit integer representation. With 8 bit images the maximum intensity value is Fmax=2^8 - 1 = 255. Pixels with intensity values around 255 are white and pixels with values near 0 are black.

Lab Topics:

Images to be used in Lab:

These images must be copied in binary format. They can be displayed in Matlab using the routines described in the Image Display section.