What is digital image processing?

Dr. Huidae Cho
Institute for Environmental and Spatial Analysis...University of North Georgia

1   What is an image?

1.1   Mathematical representation

Two-dimensional function $f(x, y)$

  • $x$ and $y$: Spatial coordinates
  • $f$: Intensity or gray level of the image at $(x, y)$

1.2   Digital image

When $x$, $y$, and $f$ are finite and discrete.

Optionally, $t$ (time) for image sequences.

A representation of a two-dimensional image as a finite set of digital values called picture elements or pixels.

What is this?

1.3   Digitization

Digitization implies approximation.

fig0215-digital-image-acquisition-process.png

Figures from Gonzalez and Woods (2002)

2   Digital image processing

2.1   Human interaction & machine perception

“Processing digital images by means of a digital computer” (Gonzalez and Woods, 2002)

Focuses on two major tasks:

  • Improvement of pictorial information for human interaction
  • Processing of image data for storage, transmission, and representation for autonomous machine perception

2.2   Three types of computerized processes

Low-level: Image input $\rightarrow$ Image output

  • Noise reduction
  • Contrast enhancement
  • Image sharpening

Mid-level: Image input $\rightarrow$ Image attributes

  • Edges
  • Contours
  • Identity

Higher-level: “Making sense” of recognized objects

2.3   Image processing vs. image analysis

Image processing: Low- and mid-level processes

Image analysis or computer vision: Higher-level process

2.4   Fields of image processing

Typical categorization

Input \ OutputImageDescription
ImageDigital Image ProcessingComputer Vision
DescriptionComputer GraphicsArtificial Intelligence

What about

  • Computation of the average intensity
  • Image compression

3   Early digital images

3.1   Digital image transfer

fig0101-1921-digital-image.png In 1921, this digital image in 5 gray levels was first sent across the Atlantic by submarine cable

fig0102-1922-digital-image.png 1922

fig0103-1929-digital-image-Pershing-and-Foch.png In 1929, 15 gray levels from London to New York

Figures from Gonzalez and Woods (2002)

3.2   First picture of the moon by a spacecraft

fig0104-Ranger7-1st-picture-of-moon.png

The first picture of the moon by the NASA Ranger 7 spacecraft on July 31, 1964

Figures from Gonzalez and Woods (2002)

4   Applications of digital image processing

  • Image enhancement
  • Industrial inspection
  • Law enforcement
  • Medical imaging
  • Astronomy
  • Object recognition
  • Augmented reality
  • Human-computer interaction (HCI)
  • Processing of remote sensing imagery
  • Etc.

5   Examples

5.1   Image enhancement

example-image-enhancement-1.png

example-image-enhancement-2.png example-image-enhancement-3.png

Figures from Gonzalez and Woods (2002)

5.2   Industrial inspection

example-industrial-inspection.png

  • Automated inspection
  • Cheaper
  • Fast
  • Reliable
Figures from Gonzalez and Woods (2002)

5.3   Law enforcement

  • Vehicle plate recognition
  • Fingerprint recognition
  • Enhancement of CCTV images

example-law-enforcement.png

Figures from Gonzalez and Woods (2002)

5.4   Human-computer interaction (HCI)

beyond-minority-report.png

5.5   Biometric authentication

biometric-authentication.png

5.6   Remote sensing

  • Manipulation of satellite imagery
  • Terrain classification
  • Landuse classification
  • Soil moisture measurement
  • Meteorology

5.7   NASA Landsat satellite bands

Band no.NameWavelength ($\mathrm{\mu m}$)Characteristics and use
1Visible blue0.45-0.52Maximum water penetration
2Visible green0.52-0.60Good for measuring plant vigor
3Visible red0.63-0.69Vegetation discrimination
4Near infrared0.76-0.90Biomass and shoreline mapping
5Middle infrared1.55-1.75Moisture content of soil and vegetation
6Thermal infrared10.4-12.5Soil moisture; thermal mapping
7Middle infrared2.08-2.35Mineral mapping
Tables from Gonzalez and Woods (2002)

5.8   NASA Landsat satellite images

fig0110-LANDSAT-satellite-images.png

Figures from Gonzalez and Woods (2002)