Re: Fractals & Pixels - what's in a name?

[Kennedy McEwen <rkm@kennedym.demon.co.uk>]

In article <4E097D008AE0D211BA2F00104B89AE900196236E@exchange1.gensec.co
m>, Arnold Theron <ArnoldT@gensec.com> writes
>The posting by Kennedy McEwen, who is involved in the technology, is very
>interesting - what especially made me take notice is the assertion that as
>many as 100 pixels could be generated from a single physical sensor - and
>that in 1976. Having developed both CCD and CMOS imaging sensors myself, I
>seriously question the idea that anything more than one pixel is obtainable
>from a corresponding physical sensor without some form of interpolation,
>even if that interpolation is done at the device level. CCD's depend on
>charge packets corresponding to the exposure, each of which need to be
>matched by a physical gate line in order to be clocked out. CMOS sensors are
>in effect individual light sensitive diodes integrated into a silicon
>substrate. There may have been other developments since, but I would really
>appreciate it if Kennedy could refer me to some published work on the topic.
>
I do hope that you are not suggesting that I am inventing, Arnold :-)
- the detail is completely OT for this list, but since you asked, and
others may question my honesty in the matter if I don't reply...  

It is a real device, though I did get the date wrong, it was originally
invented in 1974, not 1976.  It has been in production for almost 25
years and only now has the manufacturer announced an intention to cease
production in 2004 due to the capabilities finally being exceeded by
more recent developments.  Of course your interpretation of what I was
referring to may have been influenced by your own interest - I
specifically did not refer to it being CMOS or a CCD device, because it
isn't.    ;-)

Anyway, I refer you to the remarkable Transverse Electron Drift (TED)
device (an acronym many prefer to recognise as "Tom Elliot's Detector")
but also known by its trade name of "Signal PRocessing In The Element"
(SPRITE), see patents US3995159 or GB0051498/74.  Most photodetector
textbooks have a small page devoted to this invention - a few specialist
works devote entire chapters!  :-)

In this imaging sensor the photocharge generated along the full length
of strip of intrinsic photoconductor material is accumulated and swept
to an ohmic readout by the ambipolar carrier drift velocity in the
biased device.  By modifying the bias current through the photodetector,
the ambipolar drift velocity changes and, providing the carrier lifetime
of the material is sufficient, each photoconductor strip contains at any
instant the analogue signals of 50-100 pixels.  The integration time for
each pixel in the image is determined by the drift velocity and the
photoconductor strip length - so the signals are generated AND contained
separately within the sensing element.

Infrared imaging, of course, rather than visible - but in the context I
mentioned the device this distinction was irrelevant.  Each imaging
element in the sensor still contains at any instant the signals of
50-100 pixels of the image of an infrared scene.

This could only be achieved in a CCD if there were 50-100 CCD elements
to store individual photocharge packets.  The TED has no CCD, no CMOS,
no clocks, just three terminals per photoconductor.

You probably have heard about it and not recognised what it was though.
Remember the Waco seige; the FBI camera that allegedly showed the ATF
opening fire first?  That was the highest resolution, most sensitive
infrared imaging camera that FBI money could buy - TED!      ;-)

Although my original suggestion of a device containing many pixels in a
single sensor element was the TED, a Vidicon or similar tube is probably
a better and more commonly recognised example - thousands of image
pixels integrated and stored on a single element, the photocathode,
waiting to be read out by the scanning electron beam.  And that
development also predates CCDs and CMOS, going back to 1936!
-- 
Kennedy
Yes, Socrates himself is particularly missed;
A lovely little thinker, but a bugger when he's pissed.
Python Philosophers
-
Turn off HTML mail features. Keep quoted material short. Use accurate
subject lines. http://www.leben.com/lists for list instructions.