Unit DIGITAL SIGNAL PROCESSING
- Course
- Electronic engineering for the internet-of-things
- Study-unit Code
- 70A00096
- Curriculum
- In all curricula
- Teacher
- Fabrizio Frescura
- Teachers
-
- Fabrizio Frescura
- Hours
- 72 ore - Fabrizio Frescura
- CFU
- 9
- Course Regulation
- Coorte 2023
- Offered
- 2023/24
- Learning activities
- Affine/integrativa
- Area
- Attività formative affini o integrative
- Academic discipline
- ING-INF/03
- Type of study-unit
- Obbligatorio (Required)
- Type of learning activities
- Attività formativa monodisciplinare
- Language of instruction
- ITALIAN
- Contents
- The course is divided into some teaching units:
Teaching Module: Optimal and non-linear signal quantization (12 Hours)
Teaching Module: Discrete Fourier Transform and Spectral Estimation (12 Hours)
Teaching Module: z-Transforms and Transfer Functions (12 hours)
Teaching Module: Interpolation, Decimation and Rate Conversion (12 Hours)
Teaching Module: Digital Image Processing (24 Hours) - Reference texts
- S. ORFANIDIS, “INTRODUCTION TO SIGNAL PROCESSING”, PRENTICE HALL
OGE MARQUES, “PRACTICAL IMAGE AND VIDEO PROCESSING USING MATLAB”, IEEE PRESS - Educational objectives
- The course presents the basic and advanced techniques for the Digital Processing of one- and two-dimensional signals (images) and provides, together with the theoretical elements, the knowledge of some software tools (Matlab) for the application of the techniques themselves.
The course assumes that the concepts of Signal and Discrete Time System, Fourier Transform and FIR Filter have already been acquired, concepts introduced in the Signal Theory course, considered preparatory.
At the end of the course students will be able to:
Understand and design optimal and nonlinear quantization systems with design software tools (e.g. Matlab) for one- and two-dimensional signals (HDR imaging)
Understand, simulate and design (Matlab) spectral estimation schemes based on the use of DFT/FFT for both stationary and non-stationary signals.
Understand, simulate and design (Matlab) sample rate conversion schemes with direct and polyphase techniques. Applications to one- and two-dimensional signals (image rescaling)
Understand, simulate and design (Matlab) image processing algorithms. - Prerequisites
- Signal Theory
System Theory
Digital Systems Electronic
Telecommunication Basics
Measurement and Probability Theory
Computer Architecture and Operating Systems - Teaching methods
- Lectures are based on a presentation of PowerPoint Slides integrated with blackboard examples and design exercises. Interactivity with students is highly encouraged throughout the whole lecture.
For each module Matlab Exercises are carried out for theory refinement and applications. - Other information
- Learning verification modality
- The exam final test is made up if two separate tests:
1) A signal processing project on the PC (Matlab)
2) An oral exam - Extended program
- Teaching Module: Optimal and non-linear signal quantization (12 Hours)
Signal types: mono-dimensional – bi-dimensional, signal acquisition.
Fixed-point and Floating-point Numerical Formats.
Linear Quantization, Dynamic-Range and Signal to Quantization Noise Ratio (SQNR) as a function of input signal pdf
Optimal Quantization – Implementation with companding – Fully digital non-linear quantization with LUT – Applications – Matlab Exercises
Applications to Image Processing (HDR)
Teaching Module: Discrete Fourier Transform and Spectral Estimation (12 Hours)
DTFT (Discrete Time Fourier Transform) – Frequency Spectrum of Discrete Signals. Frequency Sampling of DTFT – DFT and IDFT. Fast Fourier Transform FFT.
Circular Convolution, implementation of Circular Convolution with Linear Convolution and Circular Prefix
Spectral Analysis of Deterministic Signals - DFT/FFT for spectral analysis of stationary signals- Effect of Signal Sampling - Windowing, Periodic Extension, and Extrapolation - Effect of Spectrum Sampling - Effects of Windowing: Leakage and Loss of Resolution – Design of a Spectral Estimation Software with assigned performance
DFT/FFT for spectral analysis of non-stationary signals. Spectrogram. Matlab Exercise.
Discrete Transforms for Image Processing
Teaching Module: z-Transforms and Transfer Functions (12 hours)
Main Properties. Region of Convergence (RoC), Causality and Stability. Frequency Spectrum – relation with DTFT.
Inverse Z transform
Equivalent descriptions of Discrete Time Systems, Transfer Function – Sinusoidal response, Stationary time Response, Transient Response.
IIR Filter Design with Pole-Zero placement: First Order Filters, Parametric Filters, Notch and Comb Filters. Matlab Exercise
Teaching Module: Interpolation, Decimation and Rate Conversion (12 Hours)
Interpolation and oversampling. Interpolation Filter Design – Polyphase interpolators. DAC Equalization, Multistage Interpolation.
Decimation, Decimation Filter design.
L/M Sampling rate Converters. Multistage implementation.
Matlab Exercise
Application to Image Processing
Teaching Module: Digital Image Processing (24 Hours)
Overview of Image Processing Operations: Global (Point) Operations, Neighborhood-Oriented Operations, Operations Combining Multiple Images, Operations in a Transform Domain.
Matlab Image processing Toolbox
Geometric Operations, Gray-Level Transformations, Histogram Processing, Neighborhood Processing, Frequency-Domain Filtering.
Applications: Image Restoration, Color Image Processing