Finding an asteroid is not really about seeing anything at all.
It is about noticing that one single point of light has moved while thousands of others around it have not.
The stars sit still across a sequence of exposures. A rock in the inner solar system shifts a little between frames.
That is the entire trick, and it has been the entire trick since people were doing this on glass photographic plates.
The hard part of it was never the idea itself.
It is that somebody has to look at all of it.
The work is easy and the volume is not
A single set of survey images can hold tens of thousands of separate points of light in it.
The candidate has to appear in every frame, move in a straight line, move at a plausible rate, and keep a consistent brightness while it does.
Every one of those checks is simple on its own.
Doing all of them by eye, on hundreds of objects, across a whole stack of frames, is where people stop.
The failure mode is not that volunteers are bad at it. It is that the task is repetitive, the criteria are fussy, and almost every candidate turns out to be nothing.
Which is exactly the shape of problem a computer is very good at and a tired person is not.
The rules were already written down. Nobody had put them in code for this audience.
What the program does with a stack of frames
The tool takes the image set as it comes and steps through the whole thing without being told where to look first.
It identifies the points of light, matches the same point across frames, measures how far each one travelled, and throws out everything that stayed put.
Whatever survives that gets checked against the criteria and comes back to the user as a shortlist.
The user does not read the sky. The user reads a much shorter list.
That is a change in job description rather than a change in method, and it is the difference between an evening and a month.
The author had already found two objects the slow way before writing any of it, which is where the idea came from.
The competition and the numbers behind it
The student is Siddharth Patel, in grade 7 at a public school in London, Ontario, and he was 13 at the time.
The project is called ARIA and it was entered at the national science fair held in Edmonton on May 28 and 29 of 2026.
It took the platinum award for the best junior project in the discovery category, against 390 finalists competing for close to 2 million dollars in prizes.
His own two asteroid finds came earlier, through an international search program that hands real survey images out to schools.
He had also won a gold medal at his regional fair, and first place in a dark sky photography contest the year before with an image of a comet.
The often quoted 94 percent is a drop out rate among participants in that program.
What that 94 percent does not mean
This is the number that gets repeated wrongly, and the difference between the two readings matters a lot.
Quitting a program and failing to find anything are not the same measurement at all. Plenty of people leave having found something, and plenty stay for years and find nothing at all.
A discovery rate would be a different figure entirely, and it is not the one quoted.
The award is worth stating plainly as well. It is the top junior prize in one category, not the overall prize at the fair, which went to somebody else entirely.
And a tool that shortens a shortlist is not a detection system. Confirmation still runs through professional astronomers and a formal process, in the way that credit for a first sighting or an unusual laboratory result takes years to firm up, as the fair’s own citation makes clear.
Why anybody bothers looking at all
Professional surveys sweep the sky today faster than any group of volunteers ever could, and they do it every clear night.
What volunteers still supply is follow up. An object seen once and then lost is close to useless, because an orbit needs repeated positions over time to be pinned down.
Archive images hold enormous numbers of moving dots that nobody ever remarked on, and going back through them is unglamorous work that nobody is paid to do.
That is the gap this kind of software sits in, as the coverage of the project describes it.
The telescope images were never the scarce resource here.
Attention was the scarce one.
