Title: Introduction to Grayscale and Color Images
1Introduction to Grayscale and Color Images
- Image acquisition
- Light and Electromagnetic spectrum
- Charge-Coupled Device (CCD) imaging and Bayer
Pattern (the most popular color-filter-array) - Sampling and Quantization
- Image representation
- Spatial resolution
- Bit-depth resolution
- Local neighborhood
- Block decomposition
2Electromagnetic spectrum
3Light the Visible Spectrum
- Visible range 0.43µm(violet)-0.78µm(red)
- Six bands violet, blue, green, yellow, orange,
red - The color of an object is determined by the
nature of the light reflected by the object - Monochromatic light (gray level)
- Three elements measuring chromatic light
- Radiance, luminance and brightness
4Sensor Array CCD Imaging
5Charge coupled device (CCD) image sensor
http//en.wikipedia.org/wiki/Charge-coupled_device
6(No Transcript)
7Complementary Metal Oxide Semiconductor (CMOS)
Image Sensor
http//www.dalsa.com/corp/markets/CCD_vs_CMOS.aspx
8Image Formation Model
f(x,y)i(x,y)r(x,y)n(x,y)
0ltf(x,y)lt8
Intensity proportional to energy radiated by a
physical source
0lti(x,y)lt8
illumination
0ltr(x,y)lt1
reflectance
(intrinsic images)
n(x,y)
noise
9Sampling and Quantization 1D Case
102D Sampling and Quantization
113D Visualization
It is useful to take an analogy to rain gauge
(image intensity values Measure the amount of
photon rain)
12Color Imaging Bayer Pattern
38,990
309
US3,971,065
http//en.wikipedia.org/wiki/Bayer_pattern
http//ask.metafilter.com/17138/3CCD-vs-1CCD
13Demosaicing (CFA Interpolation)
14Simple Ideas Linear Interpolation
You will be asked to try these simple ideas in
CA2
15Biological vs. Artificial Sensors
Cone distribution in human retina
US3,971,065
Question Engineers invention vs. natures
evolution, who wins?
16Digital Single-Lens Reflection (DSLR) Cameras
17Nikon D50
18Kodak Easyshare
19Photography 101
- pros
- Interchangable lens
- Greater quality and lower noise
- Suitable for high-motion and low-light
environment - Better focusing capability
- Larger focal length
- Cons
- Larger and heavier
- More expensive
- Lack of video mode
- Sensor dust problem
- More difficult to focus on very close objects
20The Plague in Photography Motion Blur
21High Dynamic Range Imaging
Q Can we generate a HDR image (16bpp) by a
standard camera? A Yes, adjust the exposure and
fuse multiple LDR images together
22HDR Display (After Toner Mapping)
Note that any commercial display devices we see
these days are NOT HDR
23(No Transcript)
24Beyond Visible
- Gamma-ray and X-ray medical and astronomical
applications - Infrared (thermal imaging) near-infrared and
far-infrared - Microwave imaging
- Radio-frequency MRI and astronomic applications
25Thermal Imaging
Operate in infrared frequency
Grayscale representation (bright pixels correlate
with high-temperature regions)
Pseudo-color representation (Human body
dispersing heat denoted by red)
26Low Signal-to-Noise (SNR) Behavior
noise
signal
27Radar Imaging
Operate in microwave frequency
Mountains in Southeast Tibet
28Synthetic Aperture Radar (SAR)
- Environmental monitoring, earth-resource mapping,
and military systems - SAR imagery must be acquired in inclement weather
and all-day-all-night. - SAR produces relatively fine azimuth resolution
that differentiates it from other radars.
29Magnetic Resonance Imaging (MRI)
Operate in radio frequency
knee
spine
head
30Basic Principle of MRI
k-space
IFT
31Comparison of Different Imaging Modalities
infrared
radio
visible
32Fluorescence Microscopy Imaging
Operate in ultraviolet frequency
normal corn
smut corn
33What Does a Neuron Look Like?
Artistic illustration
Real image
34X-ray Imaging
Operate in X-ray frequency
chest
head
35Positron Emission Tomography
Operate in gamma-ray frequency
36Mechanical Categorization of Sensors
- Motionless imaging
- Sensor is kept still during the acquisition
(e.g., CCD cameras) - Motion-aided imaging
- Sensor moves along a line or rotates around a
center during the acquisition (e.g., document
scanning and MRI scanning) - Subtle relationship between visual perception and
motion - We move because we see we see because we move
J. Gibson
37Single-sensor Imaging
38Motion Aids Imaging
39(No Transcript)
40Introduction to Grayscale Images
- Image acquisition
- Light and Electromagnetic spectrum
- Charge-Coupled Device (CCD) imaging
- Sampling and Quantization
- Image representation
- Spatial resolution
- Bit-depth resolution
- Local neighborhood
- Block decomposition
41Image Represented by a Matrix
Spatial resolution
Bit-depth resolution
42Spatial Resolution
43Image Resampling
44Towards Gigapixel
Mega-pel
Giga-pel
Photographers and artists have manually or
semi-automatically stitched hundreds of mega-pel
pictures together to demonstrate how a giga-pel
picture looks like ? the power of pixels
http//triton.tpd.tno.nl/gigazoom/Delft2.htm
45Aliasing in Digital Images
46Bit-depth Resolution
47Bit-depth Resolution (Cond)
48Commonlyused Terminology
Neighbors of a pixel p(i,j)
N8(p)(i-1,j),(i1,j),(i,j-1),(i,j1), (i-1,j-1),
(i-1,j1),(i1,j-1),(i1,j1)
N4(p)(i-1,j),(i1,j),(i,j-1),(i,j1)
Adjacency
4-adjacency p,q are 4-adjacent if p is in the
set N4(q)
8-adjacency p,q are 8-adjacent if p is in the
set N8(q)
Note that if p is in N4/8(q), then q must be also
in N4/8(p)
49Common Distance Definitions
D8 distance (checkboard distance)
D4 distance (city-block distance)
Euclidean distance (2-norm)
2
2
2
2
2
2
3
3
4
4
2
1
1
1
1
1
2
2
2
2
3
3
1
1
0
1
1
0
1
2
0
1
2
2
2
2
2
1
1
1
3
1
1
2
2
2
3
2
2
2
2
3
3
4
4
2
2
2
2
50Block-based Processing