[time-nuts] Working with SMT parts.
albertson.chris at gmail.com
Sat Aug 13 12:46:55 EDT 2016
I've been studying machine vision for years and now finally the prices are
such that I can build my own systems. There are lots of ways to construct
3D models of the environment and I think humans use many of these.
1) Using binocular vision. We humans have two eyes set some distance
apart, we all know how this works. But in humans and machines the limit,
the maximum distance is determined by visual acuity. What is the angular
resolution of the eye? Except in one tiny spot not very good Binocular
vision only work for close up.
2) Motion through space, observation of parallax.
3) observation of perspective. Perspective is the ratio of apparent size
of for grand and background object, requires familiarity with these
objects. Try it close one eye and move your thumb first in front of your
eye then at arm's length. The size of the thumb changes but nothing else
4) stereo vision by moving the head. It is easy for robot machine to take
a stereo pair of images using just one camera. It takes one then moves the
camera and tales another. Unlike the fixed human baseline the robot can
move the camera as far as it likes. Some birds do this to make up for
their short natural baseline. I think some humans try to do this too.
I'm sure there is more, especially with machines (like using more than two
cameras, sometimes a half dozen cameras or even active 3D scanners as in
I think we humans might use all of the above techniques, all at once.
Notice that we can determine the 3D position of objects on the screen of a
TV set or movie screen or even in still photographs. Not as well as with a
live view but we do manage well enough to most not be confused.
On Sat, Aug 13, 2016 at 2:33 AM, Ian Stirling <is at opus131.com> wrote:
> On 08/13/2016 12:39 AM, Chuck Harris wrote:
> > A fairly large part of the population gets along just
> > fine with mono vision. Many choose it specifically
> > by getting contact lenses where one lens is near, and
> > one is far, or getting laser eye surgery to affect that
> > condition.
> > That said, I do very well with stereo vision, but can
> > work satisfactorily well with a mono vision camera.
> > The brain is adapted to use both stereo vision, and
> > perspective for determining distance.
> Chuck, you hit the nail on the head.
> My sister, 11 months younger than I am, she was born in December 1956,
> had and still has a strange path with her vision. When she was 4,
> the eye surgeons "corrected" her turned in eye, what we called a squint.
> She is still going through life with whatever this is. When she was 32,
> she had a baby son. Later, when my nephew was 12, he was scared of the
> optical machine that the doctor was proposing to use for testing his
> eyes. She volunteered and looked into it herself.
> What she told the doctor was astonishing.
> She had no 3D vision in the sense of what most people think it is.
> For over 50 years, her brain has been constructing her 3D world, as she
> walks, as she looks around, and as she drives her car. Her 3D vision
> is based on parallax of movement. Is that inferior, as good as, or
> better than two eyes that form the 3D norm? Is your color green the
> same as mine?
> Soldering SMD parts, I use my OptiVISOR x10,
> a bit close but I can hold my breath.
> Ian, G4ICV,AB2GR
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/
> and follow the instructions there.
Redondo Beach, California
More information about the time-nuts