(^106) PART 2 HOW DIGITAL CAMERAS CAPTURE IMAGES
How the Foveon Image
Sensor Stacks It On
An obvious drawback of image sensors using
the Bayer matrix is that each pixel captures
only a third of the color. To get red, green,
and blue (RGB) values for each of the pixels,
the camera’s processor has to interpolate
the two values each pixel lacks from the
values of those colors found in adjoining light
sensors. Just as obvious, an image sensor can
gather more information from which to create
a photograph, if it can get raw RGB values
from each one of its pixels. Here’s how that
works—and how some camera makers have
come up with different solutions to the
problem.
To obtain
approximations of
the two missing color
values for one pixel, the
microprocessor runs an algorithm
on four surrounding photopixels. In the
example here, the processor adds the values
of the two green photodiodes and then divides by
- This value is joined with the values of the red and
blue to produce a virtual pixel that has values of red 200,
blue 125, and green 118. (RGB color values range from 0 to
255, with 0 representing black and 255 representing the most
intense shade of any of the three colors.)
1
The use of an area comprising four photodiodes to produce
a single pixel of roughly one-fourth that area corrupts the
resolution and accuracy of the digital image. If you have
enough pixels in an image sensor, the effect is not that obvious,
although it can still appear in moirés where fine patterns
appear in a photograph.2