This is a little overdue, but I made this sortable-searchable browser for the @Livermore_Lab YouTube films a while back, because I found YouTube's interface impossible to use to make any sense of the ~500 nuke test films they had uploaded there. Enjoy: http://blog.nuclearsecrecy.com/misc/llnlfilms/
-
Would be cool to train an AI algorithm with these videos and listed yields to automatically identify nuclear weapon yields from photos
It's not easy to do from just photos — you'd need timescales and distance references that they don't typically supply.
-
-
The yields is given too in the title. So wonder if that be used to determine distances? Or scaling? Seems like these could be used for training AI for something....
-
If you knew the timing (not always clear), yes, you could work backwards from the yield (assuming it is accurate — part of why they were re-processing this footage was to calibrate yields) to get the distance scales.
-
Currently thinking about fireball expansion rates to estimate scale..
-
You'd need to know how many frames per second the reference footage was shot in, which is not always clear.
-
Yes, that is a challenge! Many don't have frame count, and "what's a timecode?" :-)
-
And the expansion rate scales nonlinearly it seems.
End of conversation
New conversation -
-
-
Too bad none of the videos have sound.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.