[net.rec.photo] auto-focus

dalka (11/09/82)

One interesting point about auto focus cameras:

	Don't try to take pictures thru a window!!!!

				Ken Dalka
				ihuxe!dalka

hamilton (11/09/82)

#R:ihuxe:-14000:uicsovax:17200001:000:439
uicsovax!hamilton    Nov  8 23:04:00 1982

that's not really true, unless you're talking about "sonar" types
(eg polaroid).  the new auto-focus mechanisms (there quite a few now)
actually "look" at the image produced and "recognize" sharp focus.
what you have to watch out for are scenes without the edge features
that the algorithms look for.  (don't ask me for more details;
i don't fully understand how they work).
	wayne ({decvax,ucbvax,harpo}!pur-ee!uiucdcs!uicsovax!)hamilton

mem (11/10/82)

c
As I understand it, most auto-focusing mechanisms work via a
maximum contrast recognition facility.  That is, there is some measure
of the average maximum difference between light and dark spots (does
that make sense?).  (This, by the way, is in response to a reply
from wayne hamilton.)  So shooting through glass should still work,
unless the glass diffuses the incoming light.

Mark Mallett

rsx (11/10/82)

	Actually, I believe the auto-focus works on the grain of the negatives
as reflected back to the head from the print board...

	Jim Morton
	GenRad Inc.
	Concord, Mass.   (...decvax!genradbolton!grkermit!jim)

malcolm (11/22/82)

#R:ihuxe:-14000:pur-ee:8400006:000:1908
pur-ee!malcolm    Nov 10 13:35:00 1982

The lab that I work in has been looking at the technology that is 
available for measuring depth for the past few months.  We have found
three different technologies used in cameras today.  

Everybody is most familiar with the Ultrasound technology used by Polaroid.  
Their approach is to simply transmit a 1mS chirp of 50-60KHz sound waves 
and to time their return.  The timing is very simple since there is a 177 
uS delay per inch of distance.  As has been pointed out before do not try to
have the camera focus through a window.  Glass is a excellant reflector
of sound waves at this frequency.

Canon is using an infared technique in their Auto-Focus camera.  I have
been told that they scan a sharply focused infared light source through
a series of angles.  They use a receiver on the far side of the camera
to receive the scattered light.  When the light source is pointing 
directly at the object the light received by the receiver will be the
greatest and the angle of the light source can be used to estimate the 
distance to the object.  The whole thing becomes a simple triangulation
problem.  This approach also has the advantage of being more sensitive 
at closer distances (where the focusing is more difficult) because the 
angles changes faster.

I have also seen some talk of using contrast to measure sharpness of a
picture.  I believe it was somebody like Honeywell that was exploiting
this approach.  The idea is that by using three light detectors they
can form an estimate of the contrast of the image (just take the variance
of the light received at the three detectors).  It is easy to see that
the focus will be sharpest when the contrast is maximized.

I have never used any of these cameras so I can't tell you how well they
work.  I hope this answers everybody's questions.

						Malcolm Slaney
						Purdue EE Dept.
						{decvax,harpo,ihnss,ucbvax}
						!pur-ee!malcolm