Image Processing for Mechatronics Engineering For senior undergraduate students Academic Year 2017/2018, Winter Semester

Size: px

Start display at page:

Download "Image Processing for Mechatronics Engineering For senior undergraduate students Academic Year 2017/2018, Winter Semester"

Darleen Andrews
6 years ago
Views:

1 Image Processing for Mechatronics Engineering For senior undergraduate students Academic Year 2017/2018, Winter Semester Lecture 4: Fundamentals of Signal and Image Processing Dr. Mohammed Abdel-Megeed Salem Media Engineering Technology, German University in Cairo

2 Course Info - Contents 1. Introduction 2. Elementary Image Information and Operations 3. Fundamentals of Signal and Image Processing 1. Introduction (Calculus Recap) 2. Definition, 3. Important Signals 4. Signal & Image Processing 5. Sampling and Quantization 4. Image Acqusition and Digitization 5. Sensing and Perception (HVS) and the Color Image Processing 6. Image Operations 1. Point Image Operations 2. Local Image Operations and Filters 3. Global Image Operation and Transforms Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 2

3 Signal: Definition Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 3

4 Signal: Definition Important questions: What is signal Processing? But first: What is a signal? Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 4

5 Signal: Definition Important questions: What is signal Processing? But first: What is a signal? Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 5

language, image, multimedia, sonar, radar, biological, chemical, molecular, genomic, medical, musical, data, or

6 Signal: Definition By signal we may still be referring to a physical manifestation might also be symbolic or abstract information formats DNA or sequenced information Examples of signal include audio, video, speech, language, image, multimedia, sonar, radar, biological, chemical, molecular, genomic, medical, musical, data, or sequences of attributes, or numerical quanitites; Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 6

.. ist eine physikalische Darstellung einer Information... is a physical representation of information ISO/IEC 2382-1 :.

7 Signal: Definition DIN EN 60447:... ist eine sichtbare, hörbare und tastbare Anzeige, die Informationen übermittelt... a visible, audible, and tactile display for transmission of information DIN IEC :... ist eine physikalische Darstellung einer Information... is a physical representation of information ISO/IEC :... is a variation of a physical quantity used to represent data Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 7

8 Signal: Definition Oppenheim/Schafer: The term signal is generally applied to something that conveys information. Signals generally convey information about the state or behavior of a physical system, and often, signals are synthesized for the purpose of communicating information between humans or between humans and machines. Although signals can be represented in many ways, in all cases the information is contained in some pattern of variations. Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 8

9 Signal: Definition Oppenheim/Schafer: Signals are represented mathematically as functions of one or more independent 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. Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 9

10 Signal: Definition Kress "...zeitlich und örtlich veränderliche physikalische Größe, deren Parameter Nachrichten darstellen können... "... a signal is temporal and/or spatial physical variable, which may represent/carry messages in its parameters... Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 10

11 Signal: Definition Kress "...zeitlich und örtlich veränderliche physikalische Größe, deren Parameter Nachrichten darstellen können... "... a signal is temporal and/or spatial physical variable, which may represent/carry messages in its parameters... Message: information in a state of transmission, connected to communication. Information: has to do with predictability of events, is coded in parameters of a carrier Effect: elimination of uncertainty Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 11

12 Signal: Definition Information in a state of transmission: Information comes from a source/sender The processed signal is transmitted to a drain/receiver Through a channel of a physical medium Physical medium? Electric voltage or current Electromagnetic waves Sound pressure (speech) luminance (images) Info is coded in the parameters of these carrier => signals Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 12

13 Signal: Definition... a signal is temporal and/or spatial physical variable, which may represent/carry messages in its parameters... Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 13

Signal: Examples f (t) Online price-comparison site Idealo suggests consumers may be best off waiting a while to buy the S6, however, with prices expected to drop by 17% within 3 months and 28%

14 Signal: Examples f (t) Online price-comparison site Idealo suggests consumers may be best off waiting a while to buy the S6, however, with prices expected to drop by 17% within 3 months and 28% within 6 months. Audio track of singing bird. Space signal Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture

Signal: Examples f (t) Port Said Lighthouse, 1869.

15 Signal: Examples f (t) Port Said Lighthouse, Pharos of Alexandria: one of the Seven Wonders of the Ancient Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture

16 Signal: Examples f (x,y) Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 16

17 Signal: Examples Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 17

18 Signal: Examples Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 18

19 Signal: Examples f (x,t) f (x,y,t) Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 19

Signal: Examples f (x,y,z,t) http://www.linguistics.uconn.edu/asl/index.

20 Signal: Examples f (x,y,z,t) Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 20

21 Signal Signal: Dimensions Light Signal 1D Sound Rate/ Trends 2D 3D Graphics Image Video Animation Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture

22 Popup Quiz Which of the following statement regarding signals are correct? Check all that apply. A signal is a function. A signal capture the variation of some quantity or phenomena. Signals are almost everywhere. All signals are continuous. Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 22

23 Signal: Classification Not all signals are useful A simple classification of signals is into wanted vs unwanted Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 23

24 Signal: Classification Wanted signals: carrier of interesting messages, meaningful part unwanted signals: noise, affects the recognition of information White noise Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 24

25 Signal: Classification White Noise? Generated by a random process, especially important in signal processing because of its properties: Sample points are independent of each other frequency content evenly distributed arithmetic average zero Noisy Signal? is a composition of wanted and unwanted signal components: by addition or multiplication. Quantitative amount for description the portions of signal and noise: Signal to Noise Ratio (SNR) As amplitude ratio or power ratio: Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 25

26 Signal Classifications Continuous vs Discrete In terms of independent and dependent variables Multichannel & Multidimensional Signals Deterministic vs Random Signals Stationary and In-stationary Periodic and aperiodic Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 26

27 Signal: Classification let s discuss a last important pair: Deterministic signals stochastic signals Deterministic signals are described by functions. Each time at which the signal exists a numeric value may be assigned to. We have stochastic signals if each time at which the signal exists a set of possible values can be assigned to. From this set an actual value is chosen randomly. The randomness is subject to laws of probability and calculus. Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 27

28 Signal: Classification Stochastic? studying the origins of words by etymology true true sense denoting the study of sth. science of the notched not to divide ethnos (folklife studies) etymology suffix logia Entomology? a-tom Ethnology? Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 28

29 Signal: Classification Stochos: aim, target, assuming Stochastikos to be adept by guessing Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 29

30 Signal: Conclusion Signal is a carrier of information information we can get only if we have nonpredictable signals Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 30

31 Outline 3. Fundamentals of Signal and Image Processing Signals Definition Representation and Notation Important Signals Signal Processing Processing Chain Objectives Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 31

32 Representation and Notation In signal processing first: time-amplitude representation In image processing: spatial-amplitude representation In addition very important: spectral representation The combination of both: time-frequency representation Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 32

33 Representation and Notation: time-amplitude Continous Signal Discrete Signal Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 33

34 Representation and Notation: time-amplitude Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 34

35 Do not scare! It is a part of peace research Wiesel Leopard 2 Leopard 1

36 Representation and Notation: timeamplitude Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 36

37 Representation and Notation: time-amplitude Confusion matrix Typ:{} Leo 1 Leo 2 Jaguar M 48 Wiesel Leo 1 91% 0 % 0% 9% 0% Leo 2 0% 68% 4% 4% 24% Jaguar 0% 4% 77% 11% 8% M 48 0% 0% 4% 96% 0% Wiesel 0% 0% 0% 0% 100% Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 37

38 Representation and Notation In signal processing first: time-amplitude representation In image processing: spatial-amplitude representation In addition very important: spectral representation The combination of both: time-frequency representation Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 38

39 Representation and Notation: Spatial-amplitude

40 Representation and Notation In signal processing first: time-amplitude representation In image processing: spatial-amplitude representation In addition very important: spectral representation The combination of both: time-frequency representation Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 40

41 Representation and Notation: Spectral Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 41

42 Representation and Notation: Spectral Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 42

43 Representation and Notation In signal processing first: time-amplitude representation In image processing: spatial-amplitude representation In addition very important: spectral representation The combination of both: time-frequency representation Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 43

44 Time-Frequency Representation Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 44

45 Representation and Notation: t: Continuous time Notation x: Continuous spatial independent variable along the x-axis y: Continuous spatial independent variable along the x-axis f(t): Continuous function in 1D / continuous signal in time domain f(x,y): Continuous function in 2D / continuous signal in 2D spatial domain Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 45

46 Representation and Notation: Notation t n : Discrete time point. t n = n* t=nt A are equaly distributed discerete time points, t or T A is the time interval between the points (sampiling interval); n is integer. x i : Discrete spatial independent variable along the x- axis. x i =i* x is equally distant sample points. x is the x-sampling interval, i is integer. y j : Discrete spatial independent variable along the y- axis. y j =j* y is equally distant sample points. y is the y-sampling interval, j is integer. f n : Discrete function in 1D with infinite elements. f n =f(t n ) Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 46

47 Outline 3. Fundamentals of Signal and Image Processing Signals Definition Representation and Notation Important Signals Signal Processing Processing Chain Objectives Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 47

48 Important Signals: Impuls function a) Delta function, b) Periodic Delta function, c) An impulse function, and d) Periodic impulse functions. Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 48

49 Impulse Decomposition of a signal. An N point signal is broken into N components, each consisting of a single nonzero point. Important Signals: Impuls function The Scientist & Engineer's Guide to Digital Signal Processing, Smith Ch5, p101 Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 49

50 Important Signals: Sinc function for Otherwise with Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 50

51 Important Signals: Sinc function for Otherwise with Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 51

52 Important Signals: Dirac function Dirac delta (Normalized, Unit Impulse) function can be regarded as a limit value of the Gaussian function as approaches 0. Dirac delta function.. Wikipedia Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 52

53 Important Signals: Dirac you can composite every time-continuous signal as Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 53

54 Important Signals: Dirac and the time-discrete version is for Otherwise with or the composition of a time-discrete signal with Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 54

55 Important Signals: Sinc function Franciscus Vieta 1593: defined the sinc function as infinite product of cosine functions: Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 55

56 Outline 3. Fundamentals of Signal and Image Processing Signals Definition Representation and Notation Important Signals Signal Processing Processing Chain Objectives Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 56

Signal Processing: Chain process is observed and recorded physical, technical, biological process (random) initiation of actions, storage of process characteristics physical signal with meaningful

57 Signal Processing: Chain process is observed and recorded physical, technical, biological process (random) initiation of actions, storage of process characteristics physical signal with meaningful information e.g. trigger for a motor signal conversion (e.g. analog - digital) signal reconversion (e.g. digital - analog) digital electrical signal still with meaningful information meaningful information Computing, Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 57

59 Outline 3. Fundamentals of Signal and Image Processing Signals Definition Representation and Notation Important Signals Signal Processing Processing Chain Objectives Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 59

62 HOW DO WE DEFINE PROCESSING?

63 Signal Processing Proakis: A system may also be defined as a physical device/ or SW that performs an operation on a signal. Input Processing Output The system is characterized by the type of operation that it performs on the signal. For example, if the operation is linear, the system is called linear. If the operation on the signal is non linear, the system is said to be non linear, and so forth. Such operations are usually referred to as signal processing. Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 63

64 Objectives Three main objectives: Extraction of information on states or procedures of processes = extraction of characteristic values or functions Signal compression (for faster transmission or better storage) = signal compression via coding Improving of signals = noise elimination, detection of something, coloring Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 64

65 Objectives: Information Extraction Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 65

66 Objectives: Compresion Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 66

67 Objectives: Image Compression Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 67

68 Objectives: Image Enhancment a- Original image b- Laplacian image c- Sharpened image by adding a and b Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 68

69 Objectives: Image Colorization Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 69

70 Reading BRIGGS, COCHRAN, Calculus: EARLY TRANSCENDENTALS Functions, Limit, Continouty. Proakis and D. Manolakis, Digital Signal Processing: Principles, Algorithms, and Applications. 4th edition, Prentice-Hall, Signals, Systems, and Signal Processing 1.2 Classification of Signals Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 70

71 Contacts Image Processing for Mechatronics Engineering, for senior students, Winter Semester 2017 Dr. Mohammed Abdel-Megeed M. Salem Media Engineering Technology, German University in Cairo Office: C7.311 Ext Tel.: Salem, Image Processing for Mechatronics Engineering, Winter Semester 2017 Lecture 4 71

Image Processing for Mechatronics Engineering For senior undergraduate students Academic Year 2017/2018, Winter Semester

Image Processing for Mechatronics Engineering For senior undergraduate students Academic Year 2017/2018, Winter Semester Lecture 5: Fundamentals of Signal and Image Processing 23.09.2017 Dr. Mohammed Abdel-Megeed