color models
DESCRIPTION
Color Models. by Patrick Kraft Hochschule Ravensburg-Weingarten Technik | Wirtschaft | Sozialwesen Doggenriedstraße 88250 Weingarten. Contents. Introduction Physical Aspects of Colored Light Physiological Characteristics of the Human Visual System Tristimulus Color Spaces - PowerPoint PPT PresentationTRANSCRIPT
Color Models
by Patrick Kraft
Hochschule Ravensburg-WeingartenTechnik | Wirtschaft | Sozialwesen
Doggenriedstraße88250 Weingarten
Contents
• Introduction• Physical Aspects of Colored Light• Physiological Characteristics of the Human
Visual System• Tristimulus Color Spaces• Polar Coordinate Color Models• Conclusion
Introduction
Vision sensor / RGB camera is designed for the reproduction of Color for the human eye.
We first need to understand the human vision to use that kind of vision sensor.
Than we are able to convert RGB camera colors into an appropriate model for image processing.
Physical Aspects of Colored Light
Visible Light
Light and surfaces
Radiance = reflectance*radiance
Physiological Characteristics of the Human Visual System
Color vision
Normalized spectral sensivity curves of the three different types of cones
Colored light stimulates the cones and the rods
cones: s, m, l-cones
Relative spectral sensivity curves of the three different types of cones
Rods are most sensitive at 498 nm and used for night vision.
short-wavelength (s) cones 420 nm middle-wavelength (m) cones 534 nmlong-wavelength (l) cones 564 nm
There are three types of cones.
Tristimulus Color Spaces
Monochromatic Color Response
Every base color causes a certain activity on all three cones
Tristimulus space / values
We then define a given colour as
or just
CIE 1931 Standard RGB Color Space
Because of the linear algebraic properties stated by Grassmann’s laws, it is possible to represent colour stimuli by vectors in a three-dimensional space, called the tristimulus space.
CIE defined the CIE 1931 Standard RGB Colorimetric System with the monochromatic primaries:
Metamerism
Metamerism implies that two objects which appear to have exactly the same color, may have very different colors under different lighting conditions.
There are several different spectra that can appear as the same color to the observer.
Without metamerism there would be no color image reproduction on paper or screen
Bayer Array
A pattern of red, green, and blue filters on the camera image sensor.
Interpolation of surrounding colors to get the right color at each pixel
Brightness
There are twice as many green filter elements in the array as red and blue elements because: the human eye is more sensitive to green light than both red and blue light.
Therefore green colour accuracy is more important. The human eye is also more sensitive to changes in brightness than colour.
Luminance = 0.35*R + 0.55*G + 0.1*B
Polar Coordinate Color Models
Polar Coordinate Color Spaces allows movements in color space which correspond more closely to human perception
1. HueWavelength of the pure colour observed in the signal.Normalized -> Rainbow colorMore the 400 hues can be seen by the human eye.
2. SaturationPurity, a pure colour has 100% saturation
, the white and grey have 0% saturation. About 20 saturation levels are visible per hue.
3. BrightnessAmount of light emitted.Distinguishes the greylevels.The human eye perceives about 100 levels.
These spaces use a cylindrical (3D-polar) coordinate system to encode the following three psycho-visual coordinates:
HSB / HSV Color ModelTransformation from RGB <> HSV by•Travis •Foley and Van Dam•Gonzalez und Woods
Cone: Mathematically incorrect Cylinder: Mathematically correct
HSL / HLS Color ModelHue , Saturation, Lightness (also Luminance or Luminosity)
Cone: Mathematically incorrect Cylinder: Mathematically correct
IHLS Color ModelImproved Hue, Luminance and Saturation (IHLS) colour model
Saturation weighted histogram
Saturation weighted HSV histogram Saturation weighted IHLShistogram
Advantages and Disadvantages of the IHLS Color Model
Advantages
Use of real luminance: Luminance = 0.35*R + 0.55*G + 0.1*B
Saturation can be used as a continous increasing/decreasing factor on hue histograms to eleminate spikes (i.e. no jumps at black/white)
Disadvantages
expansive calculation (trigonometry) but with shader technology faster to computate
Conclusion
It is application dependent, which color model to use
There are more interesting color models used in colorimetry which are worth to study