Newsgroups: alt.ufo.reports
From: kymhorsell@gmail.com
Subject: tracking invisible objects

[uploaded 4 times; last 07/09/2024]

I've gone back to check the progress some of the long-running
programs have been making on analysing space telescope photos.  One
of the more interesting things I noted in the past was an
image-by-image scan of TESS images for moving dots.  Could we find a
series of stars across a sequence of images that were blacked out by
something moving in front of them?  That had turned out to be
easy. Yes. But surprisingly to me the length of these tracks was quite
short. It was easy to spot by eye a series of small dim stars from one
image to the next black out for 1 image then be back for the next
image but the star next to it in line had blacked out. Within a few
images all the stars were back to their pale little selves with no
sign what that blacking-out thing was or where it had gone.

So a lot hinges on what kind of definition of "object" we're giving
the AI programs. My initial mental model was something we all
understand from everyday -- some kind of pattern we can see that
maintains its shape and size and moves from one place to another;
i.e. it (partially) consistently appears a little ahead of where it
was last seen and consistently (partially) disappears a little behind
where it is now seen. It is a little tricky to get the description
right.  You should see what the programs tried to present as moving
objects for some of the definitions I tried.

But the AI's had kinda a suggestion of something to look for.  Their
ideas are based on what they can prove statistically -- i.e. what
patterns they can find in any of the numbers that appears from
examination of all the numbers over time to be very very unusual.

So what kind of patterns to try. The latest project involves spotting
a pixel in one image of a sequence.  We have to guess a heading and
speed so that in all the other images, or as many as we can find in
the collection we have, does the expected position of an object that
came from or went to (we can work backwards and forward in time since
we have all the images lined up) the original pixel appear to have
some kind of disturbance nearby at exactly the time the supposed object
passed by. Does the surrounding area suddenly become more bright or
less bright at the time the object passed through the neighbourhood.

It sounded like a good definition. But, amazingly, it manages to pick
up all kinds of things that are not easily visible in the images.
When the AI's started to parade the movies of what they found moving
around out there somewhere in the solar system it was much less than
convincing they had found anything. But after extensive arguing back
and forth it seems they are onto something. Their statistics seems to
be irrefutable. When and only when the supposed objects are near some
area it affects the brightness of stars very close to that point.  It
doesn't always affect them the exact same way. In the various "tracks"
it finds sometimes when the hidden object arrives someplace the
surrounding stars brighten a little; sometimes the same object appears
to arrive at some point and the surrounding stars dim. In many tracks
whether it dims or brightens is entirely consistent. But in many others
it is not.

On reflection this just seems to be how it should be. Some objects
passing in front of some kinds of backgrounds can made the area
appear brighter yet passing in front of other backgrounds it can make
them appear darker. A burning ember flying across the room can make
each point it arrives at appear brighter. But if there is a white
background some light from the ember that usually got lost going back
into the distance gets reflected back at the viewer. If it's a burning
ember emitting smoke it can get more complicated. If the motion
through the air fans the ember it can get even more complicated.

The main thing is -- does the supposed object being tracked "affect"
in any way the surrounding area in a way that is predictable in the
future -- i.e. in the region of space where it is moving toward.  If
so, you are tracking an object of some kind whether you can see it or
not.

The programs are presently running a "big job" to run the latest
algorithms for processing TESS "big" images to see how many objects
it can find inspecting the whole dataset.  And the dataset is massive.

The TESS instrument captures a region of the sky 24 deg by 96 deg.
TESS has 4 separate cameras of 24x24 deg lined up in a row.  Inside
each camera is a 2x2 grid of sensors (ccd's).  Each ccd is 2048x2048
"science" pixels. (Each sensor also has areas for testing the integrity
of the sensor, detecting radiation damage, and physically getting the
data out to computer memory somewhere; you can "see" all these parts of
the gadget as well in the images).  Which gives 64Mpx in all.  Each
pixel is nominally 24 bits "deep".  The machinery can take a snap every
2 seconds if so ordered.  The instrument flies around between the E and
the M so photo shoots are usually limited around 28d or less. But
that's an awful lot of pixels.

So I've been selective in my downloads to take sequences of images
from each "session" (called a "sector" in TESS parlance) 1 hr
apart. Normally some images during the month get lost because of
glare from either the E or the M or S.  There are also regular hits by
cosmic rays that can cause odd effects.  And, I suspect, some of the
images may be dropped because they show odd things. As previously
noted the number of images available on each day seems to go up and
down statistically in sync with how many UFO reports are received by
the NUFORC in the following 24-48 hrs.

But that still leaves a big number of bits to crunch. So the programs
are busy.

But here are some of the tracks it's shooting out today.  For emails
the programs have cooked up a text representation of their
results. They show a small area of some sequence of images.  Usually
the sequence is around 100 hrs long, 1 per hr.  In that sequence of
2kx2k pixels they fish out a random area typically 128x128. They cut
the 24 bits of grey scale into 3 sections -- darker than average,
average, and brighter than average which they label "-", "0" and
"+". Then for each of the tracks they find they draw a sequence of
lines showing the bright and dark pixels in the track at that given
time (i.e. from the same telescope image).

width=128 height=128 xpos=291 ypos=1313
bestsign=-1 bestlen=10 besti1=66 bestj1=4373
bestdx=0.5406 bestdy=-0.00488939
bestx1=1 besty1=34 bestx2=38 besty2=34
tracklen=37 countbest=3
00029[-]---------
00032-[-]--------
00039--[-]-------
00040---[-]------
00044----[-]-----
00065--000[-]--0-
00066--000-[-]-0-
00071--000--[-]0-
00079--000---[-]-
0009700000-0-+[-]
-37 1 34
-34 3 34
-27 6 34
-26 7 34
-22 9 34
-1 20 34
0 21 34
5 24 34
13 28 34
31 38 34
cpu=171.67 sec

The initial part of the text gives the various parameters.  Where in
the image stream and where in each image the data came from.  Which
direction the purported object was moving and where it was first
spotted. Whether it was a dark pixel or a bright pixel.  And how the
programs tracked it forwards and backwards from that point to ensure
each point the object was supposed to go suddenly got darker or
brighter at exactly that time.

The lines in the middle show the pixels in the track. It might be a bit
misleading because it does not show all the pixels around the track.
But it does show whether the track pixels are presently brighter,
average or darker than average, and it puts '[]' around the exact
pixel where the object is supposed to be at the time.

The numbers at the bottom list the pixels in the track in a simple
way.  The time before(-ve) or after(+ve) it was first spotted, the
relative X and Y inside the small section of the larger image they
were looking at.

And, of course, how long it took one of the cpu's in my small cluster
to actually do the processing. Yes. Slow. :{

But more interesting, with a growing list of these tracks we can start
to look for patterns. E.g. how fast are these objects supposedly
moving. What directions? In the case above (and in what follows,
below) were looking at "sector 55" in the TESS data.  There's a rough
plot of where that is in the sky here
<kym.massbus.org/UFO/TESS/SECTORS/s55/sect55.gif>.  
(The image shows a very rough approximation to how the TESS array
lines up against the sky).

From the latest emails I'm getting the direction the last batch of
objects appeared to be moving is:

direction= 205.442  deg
direction= 205.462  deg
direction= 205.486  deg
direction= 205.45  deg
direction= 315.52  deg
direction= 205.45  deg
direction= 25.4041  deg
direction= 25.4153  deg
direction= 95.5701  deg
direction= 205.437  deg
direction= 25.4679  deg
direction= 223.414  deg
direction= 205.45  deg
direction= 205.444  deg
direction= 110.132  deg
direction= 29.2795  deg
direction= 25.5182  deg
direction= 25.445  deg
direction= 205.655  deg
direction= 205.45  deg

So within a small 128x128 pixel area in the image stream some "dot"
that appeared to influence a series of pixels in a line was approx 205
deg was seen 10 times and another common answer was "about 25". I
immediately suspect these are related to the ecliptic. While the
AI's were supposed to try to calculate the direction as seen from
TESS with the ecliptic as the reference plane it seems some distortion
has crept in and "25" means "going right parallel to the ecliptic" and
"205" means "going left parallel to the ecliptic".

So that leaves 4 other objects spotted going at other odd angles.
That's a pretty large proportion. We know most objects in the solar
system travel around the sun in the same direction (supposedly to the
left in the TESS data I'm using here). Think a number like 99%.  A
small% go exactly in the opposite direction.  And an even
smaller number go at cock-eye angles to the ecliptic.  The known ones
are a list only 3000-4000 long. Out of many mns of known objects
down to the size of a loaf of bread. Or maybe just a large doghouse.

So it seems we are picking up things we might expect -- some kind of
rocks flying one way or the other around the sun. But we are also
picking up some odd extras.

The questions to be answered now are -- how many are we picking up in
total and is that an odd kind of number. And -- are these numbers
highly correlated or predictive of UFO sightings that are reported
around the time we're finding these rocks in the telescope image set.

I'll upload a couple movies the programs have also come up with
showing the telescope images in false colour (the TESS native
"electrons per second" that is the way it measures star brightness)
and "super false colour" where they divide the pixels just into
"average", "lighter than average" or "darker than average".  In the
movies a green line appears to show where the tracked object is
supposedly moving against the background. Although the numbers in the
data they show proves the region immediately around the supposed
object is brightening and darkening as the object arrives and departs
from each point in its track, that is mighty hard to see in the
movies.  It's the kind of thing an AI can do that evolution didn't
equip the human eye to see. Apparently rocks or strange spacecraft
flying around the solar system are not like rabbits or or other edible
items running across a grassy field.

Movie of track from some random 128x128 px region from TESS sector 55
between Aug 19 and 25, 2022. The native "brightness" as measured
in electrons per sec coming from some dot on CCD is show in a color
on the scale on the rhs. The track appears in green.
<kym.massbus.org/UFO/TESS/SECTORS/s55/movie.avi>

The movie above was "adjusted" to 3 levels, "average", "bright" and
"dark" where the average was within 100 e/sec of the average pixel
from the data.
<kym.massbus.org/UFO/TESS/SECTORS/s55/adj-movie100.avi>

Equal area split of the 3 brightness levels.
<kym.massbus.org/UFO/TESS/SECTORS/s55/adj-movie.43.avi>

--
This bill isn't all about finding little green men or flying saucers,
It's about forcing the Pentagon and federal
agencies to be transparent with the American people. I'm sick of
hearing bureaucrats telling me these things don't exist while we've
spent mns of taxpayer dollars on studying them for decades.
-- Rep. Tim Burchett

"Ridicule is not a part of the scientific method. And the public
should not be taught that it is."
- J. Allen Hynek

[REDACTED!]
US politician says she was 'Men In Blacked' as part of UFO cover-up
Daily Star, 07 Jun 2024 11:35Z
A top US politician has claimed she was blocked from investigating a UAP
report made by a USAF pilot, and hints at a wider govt cover-up of UFO...

Alameda Ends Cloud-Brightening Test, Overruling Staff Decision
NYT, 06 Jun 2024
[The experiment was testing how to spray sea salt but didn't involve
spraying it into the upper atm].

Pilot UFO Advocacy Group Forms to Offer Support to Whistle Blowers
Hosted on MSN, 02 Jun 2024
How many other relevant UFO reports have been overlooked to the detriment of
national security?

Japan lawmakers to create group for govt probes into UFOs
Japan Times, 28 May 2024
Japanese lawmakers met on Tue to create a nonpartisan group that will
ask the govt to establish an organization to investigate unidentified
aerial phenomena, or UAP, also known as UFOs. The lawmaker group will be
chaired by Yasukazu Hamada, parliamentary affairs leader of the ruling
Liberal Democratic Party.

Federal agencies face Sept. 30 deadline to share UFO data with the National
Archives
NewsNation, 18 May 2024
The federal govt's UFO information will be kept in one place by this
fall, as federal agencies must turn over all UFO-related material to the
National...

  New US law gives agencies strict instructions on releasing UFO reports
  MSN, 18 May 2024
  Federal agencies have until Oct 20th to deliver every document, audio
  and video they have about UFOs to the US govt for distribution to the...

Toxic Gaslighting: How 3M Executives Convinced a Scientist the Forever
Chemicals She Found in Human Blood Were Safe
ProPublica, 19 May 2024

Las Vegas alien video's previously missed detail proves 'authenticity,'
expert says: 'You can't deny it'
Yahoo, 10 May 2024
A veteran crime scene recreation expert said the purported UFO crash landing
in Las Vegas last year is "authentic" and he believes he spotted 2 creatures.

Retired US Colonel Predicts UFO Declassification by Oct 2030
Medium, 11 May 2024 07:35Z
Over the past weekend, the Sol Foundation, a newly formed non-profit
organization, organized a conference at Stanford University.