I don't know if this has already been worked on. But, if there was a well organized library showing video clips of things that people commonly misidentify it would be helpful. Does such a collection exist already? It would be great if each clip had a clear text description of what it looks like as well, so you can algorithmically pattern match against UFO reports.
A library of video clips showing commonly misidentified phenomena/objects would be helpful.
- Thread starter beku-mant
- Start date
Z.W. Wolf
Senior Member.
Spiderweb - single strand
The Salida Colorado UFO Sighting
https://digital.salidalibrary.org/ufo-over-salida/
The Salida Colorado UFO Sighting
https://digital.salidalibrary.org/ufo-over-salida/
Last edited:
Harabeck
Senior Member
The closest thing I've seen is the IFOs subreddit.
https://www.reddit.com/r/IFOs/
Which is sad, because it's not very active and has less than 70 posts.
https://www.reddit.com/r/IFOs/
Which is sad, because it's not very active and has less than 70 posts.
Biggest problem is every video would have to be validated and documented and provably what is described to be. Claims of 'that was really a flying saucer' would be endemic otherwise.
With Starlink videos that might not be too difficult, wait for a launch, take a crew to an identified location and wait for them to show up. Use Sitrec combined with SpaceX's launch announcement to document.
For other things... Lots of money and time and equipment.
Obtain a Police helicopter class thermal video camera setup. Hang around airports (not too close, TSA would freak out) and capture lots of airplanes arriving and leaving under various weather conditions and times of day. A couple of thousand arrivals/departures over a year or so. Document with flight schedules...
Maybe suggest AARO do this?
We've previously discussed creating a Wiki for the website, and this is another great reason for a Wiki. Unfortunately, the software this forum uses (XenForo) doesn't have a built-in ability to create one. There are third party extensions, but my recollection is each of them had significant limitations.
The motivation for me is that I am building a relatively powerful search tool that helps you find UFO reports based on similarity. You will be able to describe what you saw, and it will bring up a bunch of similarly described sightings. It is very good at pattern matching, which makes it a doubled edged sward. It makes it easy to find relevant data, but also can lead to selection and confirmation bias, because of the many spurious matches that you'll get, bias inherent in the data, and user bias based on their pre-existing beliefs.
As an example of how this can be problematic, someone recently reported an object with a yellow light at the center and green and red lights flashing around it, that they described as silent, and darting around before suddenly taking off. I ran a search on their description, and the top hit was a case from 40 or so years ago, complete with video testimony from someone who appears honest, that also involved a yellow light with red and green lights flashing around it that darted and then suddenly shot off. And according to that witness, she was later visited by people in suits who threatened her not to talk about what she saw. The recent witness thinks they saw a new kind of drone or something. But had they used this tool, and been compelled by the results, they might become paranoid, or get drawn into this alternative interpretation based on a spurious correlation.
It's clearly problematic, at least, to only return results from UFO reporting data. So, at a minimum, I think it should also return a set of possible mundane explanations, ideally grounded with examples. If the data existed, and was organized and annotated in a suitable way, the pattern matching ability could also allow you to extract the most convincing conventional examples. For example, it wouldn't just say it could be starlink, here is what starlink looks like. It would say, here is a video clip of star link from a particular perspective that creates an illusion that might explain what you specifically saw.
As an example of how this can be problematic, someone recently reported an object with a yellow light at the center and green and red lights flashing around it, that they described as silent, and darting around before suddenly taking off. I ran a search on their description, and the top hit was a case from 40 or so years ago, complete with video testimony from someone who appears honest, that also involved a yellow light with red and green lights flashing around it that darted and then suddenly shot off. And according to that witness, she was later visited by people in suits who threatened her not to talk about what she saw. The recent witness thinks they saw a new kind of drone or something. But had they used this tool, and been compelled by the results, they might become paranoid, or get drawn into this alternative interpretation based on a spurious correlation.
It's clearly problematic, at least, to only return results from UFO reporting data. So, at a minimum, I think it should also return a set of possible mundane explanations, ideally grounded with examples. If the data existed, and was organized and annotated in a suitable way, the pattern matching ability could also allow you to extract the most convincing conventional examples. For example, it wouldn't just say it could be starlink, here is what starlink looks like. It would say, here is a video clip of star link from a particular perspective that creates an illusion that might explain what you specifically saw.
Last edited:
Similar threads
