I think you’re correct, but wouldn’t this only work if you are running either android, or google maps, and have location on?
Its accurate enough but still an estimate, is the point i am getting at.
Conceivably a webcam + opencv headcount would be more precise, if the cameras covered the whole space and could account for viewing the same person from a different angle.
Its like how google can give you an estimate of bus times, but if there is a local city app that specifically interfaces directly with the actually city busses, it’ll be more accurate.
Your example with the buses is wrong. There is a standard called GTFS and public transport companies publish their fleet status and timetable according to this standard, Google just reads and displays this data. Nowadays you should see the same data in the official apps and gmaps. There are even foss solutions displaying the same thing like transportr.app
Doesn’t work that well in my experience. A place that’s mostly empty on weekdays often shows it’s really busy during weekend evenings because it is, comparatively but it’s not crowded or anything
I have this capability with my home assistant/frigate setup. Literally have a camera pointed at my back patio right now that says “Cats: 2” Cause my cats are sleeping on the couch out there.
I have no experience with the software involved in that, but I do know that generally, anything connected to a POS system probably should not be connected to a publicly accessible… anything.
Does this software even have APIs to do something like that?
Or could you just point a webcam at a screen or portion of a screen that the default software indicates open tab count on lol?
They could very easily just implement some rudimentary person identification algos and output only a headcount.
Pretty sure you can do that with OpenCV.
I think Google does this with your phone. I can see how busy various places are by looking them up on Google maps. Really useful for my local Costco.
I think you’re correct, but wouldn’t this only work if you are running either android, or google maps, and have location on?
Its accurate enough but still an estimate, is the point i am getting at.
Conceivably a webcam + opencv headcount would be more precise, if the cameras covered the whole space and could account for viewing the same person from a different angle.
Its like how google can give you an estimate of bus times, but if there is a local city app that specifically interfaces directly with the actually city busses, it’ll be more accurate.
Your example with the buses is wrong. There is a standard called GTFS and public transport companies publish their fleet status and timetable according to this standard, Google just reads and displays this data. Nowadays you should see the same data in the official apps and gmaps. There are even foss solutions displaying the same thing like transportr.app
You can browse this data worldwide on https://www.transit.land
Well dang, I did not know that.
Thank you for correcting me!
Shh don’t ruin a “viral” story
Doesn’t work that well in my experience. A place that’s mostly empty on weekdays often shows it’s really busy during weekend evenings because it is, comparatively but it’s not crowded or anything
I have this capability with my home assistant/frigate setup. Literally have a camera pointed at my back patio right now that says “Cats: 2” Cause my cats are sleeping on the couch out there.
Nice!
pretty sure they could just pull the number of open tabs
I have no experience with the software involved in that, but I do know that generally, anything connected to a POS system probably should not be connected to a publicly accessible… anything.
Does this software even have APIs to do something like that?
Or could you just point a webcam at a screen or portion of a screen that the default software indicates open tab count on lol?