Steve Fossett

Ron Wanttaja wrote:

That means you will have to zoom in on, individually, each person visible
on the image. *With average luck, you'll have to examine 200,000
individuals before you find your friend.

.... and that's why you can help with the search:

http://www.mturk.com/mturk/preview?g...G35XEZJZG21T60

You get presented with a recent satellite image of the area and can flag it
if you think it bears further investigation (note: the site is a preview,
you have to click "Accept HIT" to actually submit your opinion). If enough
people do this, we might actually manage to scan the entire area of
interest in a short time.

Ad.

--
The mail address works, but please notify me via usenet of any mail you send
to it, as it has a retention period of just a few hours.

On Sun, 09 Sep 2007 22:23:37 +0200, Adhominem wrote in :

... you can help with the search:

http://www.mturk.com/mturk/preview?g...G35XEZJZG21T60

You get presented with a recent satellite image of the area and can flag it
if you think it bears further investigation (note: the site is a preview,
you have to click "Accept HIT" to actually submit your opinion). If enough
people do this, we might actually manage to scan the entire area of
interest in a short time.

I did 102 hits. I only suggested further review on
one of them.

Had to stop a couple of times to prove I was human.
I wrote a macro to do the clicking for me.

Marty
--
Big-8 newsgroups: humanities.*, misc.*, news.*, rec.*, sci.*, soc.*, talk.*
See http://www.big-8.org for info on how to add or remove newsgroups.

"ad" == adhominem writes:
ad ... and that's why you can help with the search:

ad http://www.mturk.com/mturk/preview?g...G35XEZJZG21T60

ad You get presented with a recent satellite image of the area

Hey, are the folks organizing the Google Earth search handing out the
searchable images in a deliberate fashion? You know, so the area is
systematically viewed, not randomly.
--
If you go to a party, and you want to be the popular one at the
party, do this: Wait until no one is looking, then kick a burning
log out of the fireplace onto the carpet. Then jump on top of it
with your body and yell, "Log o' fire! Log o' fire!" I've never
done this, but I think it'd work.
- Jack Handey

On Sep 9, 12:01 pm, Ron Wanttaja wrote:
On Sun, 09 Sep 2007 10:36:35 -0400, NoneYa wrote:
We can take pictures of objects on the Earth from space that
are 2 inch's wide. We can take pictures of objects on Mars
that are 12 inches wide. Why can't we find a wrecked
airplane in Nevada?? A place that is mostly dirt and sand
with very little vegetation?

Makes no sense

No, you just have to understand the realities of the process.

Imagine a satellite snaps a picture of Wittman Field during Airventure. Assume
it has a high enough resolution to allow individuals to be recognized. There
are 400,000 people on the grounds at the time...and you want to find one
particular person. You don't know where he was at the time the photo was taken

That means you will have to zoom in on, individually, each person visible on the
image. With average luck, you'll have to examine 200,000 individuals before you
find your friend.

(Heck, here's an aerial photo of Oshkosh:

http://www.airventure.org/2007/media...al_from_SW.JPG

...just try to COUNT how many people are visible)

Keep in mind, too, that this isn't a mug shot...unless they were pre-warned, the
people in the image won't be looking at the camera. If you take the picture
from directly overhead, all you see it a bunch of caps. But even if the picture
was taken obliquely, some folks will be turned away from the camera, or holding
a cup to their mouths, blocked by other people, inside the exhibition halls, or
using a portajohn, or lying under a tree, or even unexpectedly off the grounds
entirely.

The problem is analogous to the Fossett search. Let's assume the camera gives
the equivalent of viewing an area 500 feet by 500 feet. That is about .01
square mile. With a 10,000 square mile search area, that gives one million
500x500 foot blocks to examine.

And remember all those persons who were turned away or kneeling down, tieing
their shoes, in the Oshkosh picture? After nearly two weeks of an intense air
search, the lack of success is probably because Fossett's Decathlon doesn't
strongly resemble an aircraft any more. It's undoubtedly crumpled, it's quite
possibly burned. By now, it's probably dusted with the "dirt and sand" you
refer to, making it blend in even better.

The persons who would examine the imagery wouldn't be looking for the big white
"+" of wings and fuselage, they'd be looking at every apparent bush, every
apparent rock, to guess if sometime, in the past, it just may have been an
airplane. How long should they examine each block? If each takes two minutes,
we're talking well over 30,000 labor hours. Every shadow on the image might
hide wreckage, so you'd better have another set of photos taken at a different
time of day. AND look at those.

Finally, finding hidden objects in imagery is a *military* specialty your
typical Ikonos analyst doesn't practice. If you want experts to look for the
plane, you're going to have to go to the government...and those folks are pretty
busy on some pretty important tasks.

Ron Wanttaja

There are techniques other than straight visual images for detecting
objects, although I am not sure if any of them are being used in this
context. Examples are infrared hyperspectral and polarimetric
imaging.