Digital Signal Processing
and Machine Vision
Lecture 00
Introduction
R. A. Prabhath Buddhika
B.Sc. Eng. (Hons), M.Sc.
Department of Electrical Electronic and Systems Engineering
Faculty of Engineering
NSBM Green University
Digital Signal Processing and Machine Vision
This module aims to present to students the fundamental principles and applications of
Digital Signal Processing.
The module enables students to; gain an appreciation of discrete Fourier transforms and
the relationship between them; fast computation of DFT in the context of digital filtering
and to implement (fast) filtering algorithms; analyze and design digital filters ,elementary
FIR/IIR filter design techniques, windows and bilinear transformations; to analyze the
effects of finite length words; develop basic multirate signal processing elements and
identities and to see them applied in typical applications; see, in overview, the
architectures for digital signal processing; consider the implementation aspects of simple
DSP algorithms.
2
Learning Outcomes
1. Describe and critically evaluate a range of image processing techniques
and underlying image analysis and features extraction.
2. Determine the effects of quantization and multirate digital signal
processing and image similarity matching and multi-resolution image
analysis.
3. Design IIR and FIR filters to solve basic filtering problems. Analyse the
spectrum of Discrete-time signals using the Discrete Fourier
transformation.
4. Design and implement solutions using MATLAB for common DSP
problems and building an intelligent machine vision system.
3
Assessment Criteria
An Exam weighted at 60% and Coursework weighted at 40%.
Coursework: In-class test weighted 20% (LO1 and LO2) and Lab report
weighted 20% (LO4).
Exam: Final examination length 3 hours covering all LOs
4
Tools to be Used
MATLAB
Octave & Open-CV
5
Learning Resources
1. Oppenheim et al. 2010. Discrete Time Signal Processing. 3rd Ed. Pearson.
2. Gonzalez et al. 2018. Digital Image Processing. 4th Ed. Pearson.
6
Course Management
https://nlearn.nsbm.ac.lk/
7
Semester Plan
We will discuss Digital Signal Processing in the first ⅔ of the semester.
Machine Vision will be discussed in the last ⅓ of the semester.
That is because we will need the concepts of DSP to understand machine
vision.
8
What is a Signal?
The term signal is generally applied to something that conveys information.
Signals may, for example. convey information about the state or behavior of a
physical system.
-OPPENHEIM and SCHAFER
Signals are represented mathematically as functions of one or more in
dependent variables. For example, a speech signal is represented
mathematically as a function of time, and a photographic image is represented
as a brightness function of two spatial variables.
9
What is Digital Signal Processing?
Digital Signal Processing(DSP) is the branch of electronic engineering
that studies, representation of signals in digital form and analyzing,
modifying and extracting information from them.
Engineers use a special class of processors called Digital Signal
Processors to implement the methods to manipulate digital signals.
These processors are expected to deliver relatime performance.
10
Digital Signal Processing
In DSP we often consider signals as a function of an independent variable and
the measurable quantity of the signal is the dependant variable.
In practise, most common independent variable is time measured in seconds
and the most common dependant variable is the amplitude measured in
volts(in some special cases, current in Amperes).
There can be other quantities also taking the place of time and/or amplitude.
However our main focus in this module are on single independent variable,
time, and single dependent variable, amplitude.
11
Example
12
Different perceptions of signals
1. Continuous time - Continuous amplitude
2. Continuous time - Discrete amplitude
3. Discrete time - Continuous amplitude
4. Discrete time - Discrete amplitude
13
Different perceptions of signals continued…
14
Different perceptions of signals continued…
Continuous time - Continuous amplitude signals are
called Analog Signals.
Discrete time - Discrete amplitude signals are called
Digital Signals.
15
Different perceptions of signals continued…
ANALOG
DIGITAL
16
Real World Problem
Natural signals are mostly Analog!
How do we deal with Analog signals using Digital computers???
We convert Analog signals to Digital in three steps.
1. Sampling - Discretize the Time axis
2. Quantizing - Discretize the Amplitude axis
3. Encoding - Convert the quantized sample values to a convenient way of representation.
Analog to Digital Converters do this.
A reverse conversion is done by the Digital to Analog converters.
In DSP we need both these to interact with real world systems.
17
