Applications of Image Processing

1

• Applications of Image Processing
• General Requirements
• Selected Applications

2
Faculty Dr. Lester A. Gerhardt Associate Dean,
School of Engineering Rensselaer Polytechnic
Institute Phone [phone] Fax
518-276-8788 E-mail gerhal_at_rpi.edu
3
Typical System Requirements forDigital Image
Processing Applications

• Image Capture
• Sensors / Cameras
• Optics

• High Speed I/O
• Converters

• Appropriate Display Capability

• Large and Fast Memory

4
Typical System Requirements forDigital Image
Processing Applications

• Software
• Operating System Software
• Applications Software
• User Friendly and Interactive

• Performance Evaluation Metrics
• Human Observation
• Machine Observation

5
Applications Driven Imagery Requirements

• Static / Dynamic Imagery

• Speed - Real Time / On Line

• Resolution

• Sequential Image Processing

• Image Comparison

• Image Reconstruction

• Multi-source Data Fusion

6
Applications Driven Imagery Requirements

• Integrate with Design Data Bases

• Storage and Retrieval

• Communications Requirements

• Interfacing

• Lighting

• Ground Truth

7
IMAGE PROCESSINGSome Diversified Applications

• Space
• Eyes in the Sky
• Weather, Earth Resources,
  Military, GPS
• Planetary Exploration

• Medical
• Tomography (CAT), PET, MRI, Ultrasound,
  
• X-Ray, SPECT
• Phonocardiogram

8

• Biometrics
• Handprints -INS
• Fingerprints
• Voiceprints
• Dynamic signatures
• Retinal eye maps
• Iris scanning

9

• Inspection
• Model Based and Object Based
• ( Sampling / View Planning
  )
• Sandpaper Production
• ( Process Parameters
  Control )
• Automotive Locking Nuts
• ( Classification )
• Glass Bottles
• ( Defects )
• Models From Objects
• ( Reconstruction )
• Roadway Inspection
• On Line PCBs
• Aluminum Rolls

10

• Eye Tracking
• Patent - January 2, 1996
• Aid to the Disabled
• Cameras - 35mm and Camcorder

• Other
• Art ( Forgeries, Comparative Analysis )
• Facial Expressions
• Food Processing
• Traffic Management

11
Face recognition
12
Examples of Dominant System Requirements for
Selected Applications

• SPACE
• Communications
• Limited Power and
  Bandwidth
• Error Correcting Coding
• Resolution
• Image Motion Compensation
• Image Registration
• Geometric Distortion Correction
• Sensors

13

• MEDICAL
• Sensors ( Spectral, Spatial, Temporal )
• Image Construction
• Image Storage and Retrieval
• Image Registration and Comparison
• Image Analysis and Enhancement

• Fingerprints
• Sensing and Image Quality
• Image Enhancement
• Feature Extraction
• Registration / Matching

14
SPACE APPLICATIONS
15
Imagery From Space

• Orbital
• 80 to 120 miles
• Geo-stationary
• 22,000 miles
• Moon (1969)
• 240,000 miles
• Uranus (1986)
• 1.8 billion miles
• Neptune (1989)
• 3 billion miles

16
(No Transcript)
17
These two pictures of Uranus-- one in true
color(left) and the other in false color

• Take Jan. 17, 1986

18
Miranda, innermost of Uranus large satellites

• Taken Jan. 24, 1986

19
Image Motion Compensation(MathWorks Image
Deblurring)
20
(No Transcript)
21
(No Transcript)
22
MEDICAL APPLICATIONS
23
REFERENCES

• SPECIAL ISSUE ON
• MEDICAL IMAGING MODALITIES
• IEEE SIGNAL PROCESSING
• JANUARY 1997
• VOL 14, NO. 1

24
Conventional Radiography
25
Conventional Tomography
26
MRI
27
Positron-Emission Tomography
28
Ultrasound Image
29
BIOMETRICS
30
Biometric man
31
Comparison of Biometric Technologies
32
BIOMETRICS REFERENCES
THE ART and SCIENCE of IDENTITY
VERIFICATION IEEE SPECTRUM FEBRUARY, 1994
AUTOMATED BIOMETRIC SYSTEMS IEEE
PROCEEDINGS SEPTEMBER, 1997
33
Examples of Personnel Verification and
Identification

• INSPASS
• Hand Geometry

• SENSAR ( Moorestown, N.J.)
• Iris Scanning
• ATM Application
• Estimated to Cost Less Than 2,000
  

34

• PRINTRAK, STARTEK, CAMBRIDGE NEURODYNAMICS
• Fingerprints

• RENSSELAER
• Multi Modal Sensing
• Fingerprints
• Voice
• Dynamic Signatures

35
Your hand is your passport
36
BioMouse
37
Fingerprint Verifier by Startek
38
Gun Grip Microchip Reads Fingerprints(Oxford
Micro Devices)
39
Classical Fingerprinting

• Card Format
• 10 Prints / Card
• Inked Rolled Prints - Nail to Nail
• Cooperative Client
• High System Cost
• Large Data Base
• Reasonable Error Rates
• Somewhat Controllable Environment

Who is This?
IDENTIFICATION
40
(No Transcript)
41
DIRECT READING
Real Time on Line 2-5 Seconds Contact Methods (
DAB Print ) Noncontact Cooperative User Low
System Cost Extremely Low Error Rates Required
(Type I and II) Environmentally Variable
Is This Mr. Chen?
VERIFICATION
42
Personnel Verification Using Fingerprints
Dominant Functions

• Image Capture
• DAB Fingerprint
• FTIR Prism
• Display

• Image Enhancement
• Noise Removal
• Line Thinning
• Ridge Tracking

43

• Feature Extraction
• Global Features
• Core, Delta, Ridge Count, Type
• Minutiae Location
  

• Matching
• Verification

44
3-D Fingerprint Image
45
A Complete Tracing of An Ulnar Loop Pattern
46
General Classification
Tented arch Loop
Plain whorl Double loop
47
Ridge counting
12 counts 2 counts
16 counts 7 counts
48
Automatic Core Locating
49
Fourier fingerprinting technique
50
(No Transcript)
51
Inspection Applications
52
INSPECTION

• Model Based / Object Based
• Sampling ( Non Uniform )
• View Planning

• Sorting Automotive Locking Nuts

• Glass Bottle Inspection

• Model Reconstruction

• Roadway Inspection

• On Line PCBs

• Aluminum Roll Flaws

53
(No Transcript)
54
INSPECTING STATE ROADWAYS

• Sensors
• Digital Video Cameras ( Sony DX 9000 )
• Lasers
• Ultrasound ( 56 )
• GPS
• DME

• Use of Strobe Lighting

• Ruts as Small as 100mm

• Cracks - 3mm Wide

55
Automatic Road Analyzer Connecticut Department of
Transportation
56
On LINE PCB INSPECTION

• 50,000 Coordinates in 0.1ms

• Measurement Accuracy Better
• Than .001 FOV

• PCBs from 4x4 to 18x18 Inches

• Intelligent Automation Systems
• ( Cambridge, MA )

57
INSPECTING / DETECTING

• Raw Material
• Rolls .005 to .02 Inches Wide
• 100 Feet / min Web System

• Sensing
• Four 2048x1 8 Bit CCD Cameras
• Mayan Automation ( Canada )
• .005 Inch Scratches / Defects

58
EYE TRACKING
59
(No Transcript)
60
Canons new eye controlled focus is operated by
the movement of your eye.
61
OTHER
62
FOOD PROCESSING

• National Frozen Foods ( Burlington, WA )
• 60 Million Pounds of Vegetables / yr
• 20 Million Pounds of Peas

• Machine Inspection and Sorting
• Key Technology ( Walla Walla, WA )

• PEAS
• Sorted Solely by Color ( Table Look Up
  )
• Graded / Separated

63

• System
• Four Cameras ( 2 Above - 2 Below )
• 2048 Pixels
• 10 Bits / Channel
• 60 Inch Wide Conveyor
• 590 Ft/ min
• 30 Frames / sec

64
TRAFFIC MANAGEMENT

• Intelligent Highway Vehicle System

• Intelligent Transportation Infrastructure

• Advanced Traffic Management System
• (1995-2000)
• Video Image Processing - 15 of Project

• Sensors
• Solid State CCTV Video
• Wide Area Detection (Six Lanes)
• Autoscope System
• (Econolite Control Products - Anaheim,
  CA)

65
(No Transcript)
66
Robotics and Vision - Examples

• Pick and Place Part Inspection
• SEIKO D-TRAN System Robot

• Movable Color CCD Cameras
• University of Rochester
• Dual Eye Platform
• 400º/sec Possible ( Human Eye - 700
  º/sec )

• 3-D Tubular Objects Bin Picking
• Purdue University

67

• Three Camera Ranging Platform
• Carnegie Mellon University
• 20 MFLOPS

• COG Robot
• Approximates Human Eye Movement
• Peripheral Vision and Foveal Vision

68
CenSSIS Image Processing Examples
69
Electrical Impedance ImagingLung and Heart
Phantom
70
Electrical Impedance ImagingReconstructed
Resistivity Image
71
Presentation Team
Badri
Hanu
Chuck
72
General Definitions
73
Spatial Mapping
Mosaic
74
Example Application Underwater Mosaics
75
Global Image Characteristics
Fourier Transforms

• Example translation in image plane becomes
  phase-shift in frequency domain
• Rigid transformations, scaling
• Large rotations

Wreckage from MV Derbyshire 4200m depth, Pacific
Ocean
76
Issues and Results
Whale Remains 1700m water off San Diego (15m
across)

• Intensity differences may be accommodated by
  normalization or by smoothing and differentiation
• Commercial products are available