Hey Google, is serving a lower-resolution image to Firefox part of leveraging a monopoly against Firefox?
Is your legal team fine with that? Is there any legitimate reason anyone could possibly identify for why you would purposefully worsen the experience for a fully compliant browser like Firefox, by lowering the resolution?
Google has two sets of base images. I usually prefer the plain 2D images, but you can get their reconstructed/interpolated 3D view with false color by toggling the "Globe" option in Google Maps, or the "3D Buildings" option in Google Earth. At this particular location, the 3D view has higher resolution.
The single thread WASM beta version appears to work perfectly on Waterfox (the multithread version loads data then sits there doing nothing). Not the fastest thing around but useable and noticeably better than maps.
No, Google Maps is actually a case where JavaScript makes sense. But this message reminded me of many seemingly static pages that are actually blank if you don't enable JavaScript, without even any hint that you should enable JavaScript. Reader mode often works perfectly on these pages. Sometimes, blocking the white element hiding the content of the page works well too.
I don't hate JavaScript, I have actually developed several web applications using it, and like it.
The 3D view looks so comically dystopian. The distorted effect makes everything look mangled.
On the other side of the pond, about 4 houses down... the backyard looks burned out and the pool is drained and bloodstained. This is like /r/writingprompt/
What is up with that red spot in that empty pool for real though? There's a truck parked out front with the door open, obviously people are working on the back yard in general, but what in the world would that red spot even be? Bing maps shows the backyard done and the pooled filled and normal-blue-looking.
In some areas the soil has a lot of iron in it and looks reddish (like rust-colored literally). But I'm not sure whether this is one of those areas, and also the particular shade of red here seems a bit more vivid than just rusty-red!
It could just be a tarp or a toy that melted. Pools always attract a lot of dirt and insects. Without constant maintenance it will look discoloured pretty quickly.
There's also construction-looking vehicles parked out front of these house so it's probably under renovation.
FWIW Google Earth VR on PC does let you walk around the 3D view and you can set your scale to be real-world. I'm assuming it's the same 3D reconstruction that you get on regular Google Maps / Google Earth.
You can turn off 3d mode by the blindingly obvious procedure of clicking the hamburger button, observing that "Globe" mode is enabled (because it is blue) and then clicking "Globe" to disable this mode.
Don't be so foolish as to think the button labelled "3D" in the content area toggles the 3D mode. Only an idiot would make that assumption!
(The above is sarcastic. I think the discoverability of this feature, and the weird 1/3/1 modality behavior of the items in the second section of the hamburger menu is shockingly poor).
The quality drops off if you do that and you get a completely different set of images. There's no way to turn of the 3D garbage and still see the same thing.
Nevermind it being visible on google maps since 2007...
Imagine being the homeowner who had no idea a corpse was decomposing steps from his own backyard for the past 22 years... or that a car was in the pond behind your house.
Could you imagine going out for a simple swim, diving down for fun, and stumbling across a car with a rotting corpse in it? There's the reason why I don't like swimming in deep water.
It's a man made pond in Florida, it's not deep at all. Definitely not something you'd swim in though, lots of funky stuff grows in warm stagnant water like that, plus ever present alligators.
Visiting friends in Florida recently, they warned our kids not to go in the landscaped pond in their development (it had a beach). I asked if snakes were a danger. They laughed, not in the least: the real problem was flesh-eating amoebae.
Ugh I thought that was amusing, at least until I read yesterday about a 10 year old girl in Texas who died from exactly this.
The amoeba is called Naegleria Fowleri and nagelerisis is a terrifying way to die: even with treatment there's a 95% mortality rate.
The amoebas enter the nose when they traverse the olfactory nerve via something called the cribriform plate to enter the brain. Once there, they mistake neurons for bacteria, which they consume.
As someone who used to be a Floridian, tales of gators in ponds are greatly exaggerated :) The vast majority of ponds, especially community ponds, do not have gators.
Florida has unusual issues. The groundwater isn't very far below the surface and you basically can't have basements there. Some parts of it are swamp and it has a lot of karst that is prone to sink holes.
(Off the top of my head -- so anyone more knowledge, feel free to nitpick that to death. Alternately, anyone desperate to pick a fight and not knowledgeable enough, feel free to google my remark to within an inch of its life.)
Native Floridian. I spent my youth shoeless around these ponds, catching and releasing gators, snakes, and fish. I find those experiences growing up invaluable.
Inland coastal Alabama, same deal. It's great to know how to catch turtles, snakes and lizards with my bare hands. Doesn't come up much while developing software but it definitely balanced out the Super Nintendo time growing up.
I’ve done other things that almost got me killed but nothing gives me the heebeejeebees like thinking about the time I drunkenly peed into a river delta near daintree. I don’t even want to think about what was waiting for me to step closer to the water...
Native Floridian here as well. Same, but I would never swim down into a submerged car to investigate it. They're usually far too disgusting and too deep down. There was a submerged car in a canal where I grew up, I used to use it as a landmark for fishing all the time. It's probably still there, who knows, there could be a body.
If he lived on that street or a connecting street then you can see how he might drive right into that pond if he fell asleep (from all those drinks at the club) right before making that turn.
I thought about that as well. I guess there are any number of answers and maybe they did check the ponds and couldn't see anything -- it was visible on google maps for 12 years and nobody spotted it so it couldn't have been that easy to see from the shore.
Other comments reveal that, in fact, this is inside a gated community. Driving your car into a gated community and then into a pond just doesn't add up as an accident.
Ah, OK; well that paints a bit of a different picture: driving through some construction site which might have been in who knows what sort of disarray at the time, without proper roads and hazards/obstacles and whatnot.
It probably wasn't visible from ground level. If it's five feet below the surface but 30 feet from shore, then your line of sight would pass through a bit more than 30 feet of water, and the pond is muddy enough that it's partly obscured even from directly above. If no one happened to swim or boat directly over it, and with no tall buildings nearby, there may have been no opportunity to see it.
(I've noticed something similar flying over the Great Lakes. Looking down from the window of a descending plane, you can clearly make out lakefloor contours that are invisible from shore, even with the relatively clear water.)
Also the Fresnel reflection coefficient comes into play [0]: water surface has higher reflectivity at grazing-incidence angles than at those close to the surface normal.
When the vehicle was pulled from the water, skeletal remains were found inside. One week later the remains were positively identified as belonging to Mr Moldt.
Decomposing is typically only in reference to tissues/organs/soft matter. Things that bacteria, insects, and animals can eat (thus causing the decomposition). This happens rather quickly.
Bone can't be eaten, so when all that is left is the Skeleton there is no more decoposing. As long as they aren't crushed or destroyed through other means, bones will last virtually forever.
I assume he was identified by dental records or something similar (healed broken bones, etc).
>Palm Beach County Sheriff's Office told the BBC that Mr Moldt is presumed to have lost control of his vehicle and driven into the pond.
>The force said that, during the initial investigation into his disappearance, there was "no evidence of that occurring" until recently, when a shift in the water made the car visible.
I'm surprised that no one saw signs of a car going off the road back when he went missing. In the image the car is quite far from the road suggesting a lot of momentum as it went into the pond.
I wonder how hard it would be to scan ingested maps for interesting submerged structures, and also roughly how many cars would turn up if it's feasible. I'd guess... O(100)?
Space-based radar is also available which provides views of metal content as an intelligence product. Brighter parts of the image = more metal. I'm not sure if it's commercially available but it would sure help to find cars in shallow water...
There is a lot of other metal in shallow water though. Just think about rubbish, buoys, marker posts for utilities, boats, sewer outfalls, gratings etc.
I can't find a link to the story at the moment (I'm on mobile), but I read recently they've started doing just that - doing image analyses on aerial/sat map data - to find cars and other things submerged in the Bayous here in Houston. Previously [1] they used more on-site resources, such as sonar, drones, and people in kayaks; etc.
I found a skateboard in a smaller bayou this past weekend in fairly good condition. I can’t imagine what else is down there. There is also a very surprising amount of fish in the bayous, including loads of plecos (aquarium sucker fish).
I wonder if there's anyone scanning satellite imagery for the word "HELP", or if we should agree on some other symbol people could use to ask for help that would be easier to discover with computer vision and wouldn't be English-specific.
How large should you make the letters? Satellite resolutions range from 2m for Planet Labs, 50cm for Geoeye to ~10cm for that spy satellite Trump leaked recently. The MNIST dataset of hand written digits uses 28*28 pixels, but you can get away with less. Maybe 50 meter by 50 meter letters?
If you know binary you could also include extra information by using rows of 5 squares ~1m in diameter.
Google Maps uses aerial imagery (photos taken from planes)[0] because it's higher resolution, and I bet that's what you're seeing here. They don't fly those planes often and they don't fly them over random uninhabited islands, whereas there's a company that takes satellite images of all the land on the planet daily.
I can’t imagine getting the entire world to learn a new symbol for “help” on the off chance someone needs to signal to passing satellites. Also, I believe SOS is fairly well-known.
How frequently do we refresh our coverage of imagery of planet earth?
Some sort of universal symbol could be great. Especially if you knew that within X days there was a strong chance it would be spotted and rescue folks alerted.
f(x) is O(1) means that there exists x0, M > 0 such that for all x > x0, |f(x)| <= M. Clearly, the statement still holds (for different M) if 1 is replaced by any positive number.
also visible in Apple Maps. Definitely in 3D, but only just in 2D. I.e. in 2D at maximum zoom there is a light spot where the car was. 3D lets you zoom in closer, and you can see the rear end of a car even more clearly than the picture in the article.
If you go to https://www.historicaerials.com/viewer [1] their resolution isn't great (can't see the car), but you can see that in 1999 (~2 years after his disappearance) they were still building houses there... ~2 years before (1995) that area was orchards or farmland.
[1]I used 3789 Moon Bay Cir
Wellington, FL 33414
United States
Last month, I noticed there was a sign not well fixed over the road. So when I arrived home, I reported it to the appropriate authorities through a web interface. And doing so, to pinpoint it for the repair guys, I noticed that the precise sign was available to be seen via street-view. I was even able using the time-travel feature to notice that it has been dangling for more than two years.
Probably running anomaly detection on all road-signs all over the world, you could prevent a certain number of accident each year.
The deeper question behind this, is what should have been done by those who have the info and didn't act ? Once you start running anomaly detection on data how do you navigate the ethical minefield that ensue.
Well depending on the pricing of the API there are plenty of low revenue services that can't be run profitably.
So depending on the pricing of the API, you let accidents happen. Or even worse, depending on whether or not the municipality is rich enough to buy your services you let them happen or not (even though the poorer the city the more likely it is that they already have maintenance issues).
There are plenty other things you may notice on street-view,
like dangerous/illegal behaviors, pot-holes, even finding cars in lake can have an impact of property value.
Reporting anything is probably a legal or PR risk because of the unknown ramifications so you end-up doing nothing and the accidents happens, yet no-one can hope to offer a better service because it doesn't make economical sense because of the existing service tailored to another usage of the data.
That's not an easy problem but that's one which is at the heart of all-data related business. For example, when you have access to plenty of face pictures with flash you may notice some medical issues with eyes, obviously this is a valuable information for the health insurer, and also to the user.
A friend interned with the Wisconsin DOT. One of his jobs was to review footage from what was earnestly "Google Streetview with higher resolution".
Actually Tesla might have a good dataset to work with since their system relies heavily on vision. Imagine location, timestamp, and sign verbiage sent to a server, if it stops seeing the sign in that area, also return an image next time a Tesla drives by for manual review, and file a maintenance request with the DOT with the click of a button.
> "Amazingly, a vehicle had plainly [been] visible on a Google Earth satellite photo of the area since 2007, but apparently no-one had noticed it until 2019," according to the report.
This sounds like a job for machine learning. Are there any systematic efforts underway to use Google Earth to search for out-of-place objects?
That's where Microsoft's old TerraServer got their stuff, too, IIRC, which is the first place I remember seeing anything like that online. Wasn't that much worse than Google's version, really, considering how much earlier it was.
I think my parents had an aerial photo of their land a few years before even TerraServer was came around, from some guy who'd flown over a big chunk of territory taking photos and then sold them to the land owners. I assume this was far from the only person doing that, all over the US. Pretty cool back then, when a clear shot from the air of the area you lived in wasn't something you could grab at a computer any time you wanted, or with a cheapish drone and a few spare minutes. And before everyone else could creep on your space in the same manner on a whim, for that matter.
I don't know if "most" is totally accurate, but definitely the highest resolution ones and probably the most recent ones. Probably anything less than 1m.
There is a significant discrepancy between the imagery viewable in 3D mode and the stuff that's only 2D, because Google simultaneously captures high resolution aerial photography and Lidar, particularly around cities. You can see vast differences in urban or rural landscape sometimes because the stuff you see in Google maps can be taken as far as years apart from the 3D aerial stuff.
For example, you can see where the data source and rendering method change just outside the city of Winnipeg on the South end. Probably testable in many other cities simply by seitching between google maps and google earth.
This same story has happened before, in Michigan, when a worker putting christmas lights on a nearby tree noticed a car in a pond from their raised vantage point.
I wonder how many other cases like this can be solved with an image recognition based AI scraping the rest of Google Maps out there.. anyone on here interested in collab on that?
I've found and reported obviously problematic submerged items on Google Maps on a few occasions. No one is interested. Now I just share them with my friends.
you should post them on a blog or make a subreddit or something. This comment section would seem to indicate people love looking at weird stuff on Google Maps.
So if a whole car can stay hidden for 22 years, that tells me that in those 22 years, nobody swam or fished nearby. The aerial photo makes it obvious the whole subdivision is intercut with ponds that are obviously man-made, and they look stagnant, nasty & sad. Probably also full of pesticide & herbicide runoff(t). Everybody in the subdivision probably tells their kids to stay out of the ponds. "That's what the pool is for honey! Ponds are for poor people!" I wish Kate Wagner (mcmansionhell.com) was here.
They don't look that stagnant to me? Each one seems to have multiple inlet/outlet pipes, and the next pond over has an aerating fountain. If they were truly stagnant the car wouldn't be visible.
201 comments
[ 3.4 ms ] story [ 225 ms ] thread[mobile] https://www.google.com/maps/@26.6249928,-80.227623,23m/data=...
-- thanks to comment below for mobile link.
The Chrome one seems to be much higher quality.
Is your legal team fine with that? Is there any legitimate reason anyone could possibly identify for why you would purposefully worsen the experience for a fully compliant browser like Firefox, by lowering the resolution?
What does your legal team have to say about this?
Ah ah! Quite incredibly and sadly spot on.
I don't hate JavaScript, I have actually developed several web applications using it, and like it.
Google is the worst.
Aw snap! Google Earth isn't supported by your browser yet. Try this link in Chrome instead. If you don't have Chrome installed, download it here.
Or if you're feeling adventurous, you can preview a cross-browser beta of Earth built on WebAssembly.
On the other side of the pond, about 4 houses down... the backyard looks burned out and the pool is drained and bloodstained. This is like /r/writingprompt/
There's also construction-looking vehicles parked out front of these house so it's probably under renovation.
It's definitely a must-try VR application.
Everyone I demoed the first Vive to loved it, even though the reconstruction is far from perfect.
sightseeing (1:1)
you may be a nobody in real life but you and your clan online control the Rome Colosseum
[0]: https://vr.google.com/earth/
[1]: https://cloud.google.com/maps-platform/gaming/
Don't be so foolish as to think the button labelled "3D" in the content area toggles the 3D mode. Only an idiot would make that assumption!
(The above is sarcastic. I think the discoverability of this feature, and the weird 1/3/1 modality behavior of the items in the second section of the hamburger menu is shockingly poor).
It can, surely, if you want it to?
Imagine being the homeowner who had no idea a corpse was decomposing steps from his own backyard for the past 22 years... or that a car was in the pond behind your house.
People really need to get out more often.
The amoeba is called Naegleria Fowleri and nagelerisis is a terrifying way to die: even with treatment there's a 95% mortality rate.
The amoebas enter the nose when they traverse the olfactory nerve via something called the cribriform plate to enter the brain. Once there, they mistake neurons for bacteria, which they consume.
Snakes? Another story.
The need for this qualification on your statement really isn't helping. /s
Yea, local sterotypes tend to be that way, Boston is actually a really nice city with (mostly, except the crazy insulated catholics) pleasant people.
People want to live there, so they actively and continuously drain the swamp and build wherever possible.
Sea level rise is posing some expensive problems for them, obviously.
2) use fill dirt to raise land for houses so they are above flood levels
3) sell houses with "waterfront views"
(Off the top of my head -- so anyone more knowledge, feel free to nitpick that to death. Alternately, anyone desperate to pick a fight and not knowledgeable enough, feel free to google my remark to within an inch of its life.)
One man's trash is another man's treasure.
https://en.wikipedia.org/wiki/Freedom_of_information_legisla...
https://www.miaminewtimes.com/news/how-floridas-proud-open-g...
(I've noticed something similar flying over the Great Lakes. Looking down from the window of a descending plane, you can clearly make out lakefloor contours that are invisible from shore, even with the relatively clear water.)
[0] https://en.wikipedia.org/wiki/Fresnel_equations
A corpse is not going to decompose over the span of 22 years; it's pretty much done within the first year.
When the vehicle was pulled from the water, skeletal remains were found inside. One week later the remains were positively identified as belonging to Mr Moldt.
Bone can't be eaten, so when all that is left is the Skeleton there is no more decoposing. As long as they aren't crushed or destroyed through other means, bones will last virtually forever.
I assume he was identified by dental records or something similar (healed broken bones, etc).
Ironically, like plastic. :)
Squirrels eat bones. Skeletons disappear rapidly in the wild.
>The force said that, during the initial investigation into his disappearance, there was "no evidence of that occurring" until recently, when a shift in the water made the car visible.
I'm surprised that no one saw signs of a car going off the road back when he went missing. In the image the car is quite far from the road suggesting a lot of momentum as it went into the pond.
Tire tracks in grass (if he tried to stop at all) wouldn't be as out of place.
I bet IR scans could help, similar to how they are discovering ancient cities buried deep in the jungle.
So any companies make this kind of imagery data available free; I'd be interested to see some.
1: https://www.houstonchronicle.com/local/gray-matters/article/...
https://en.m.wikipedia.org/wiki/Magnet_fishing
How large should you make the letters? Satellite resolutions range from 2m for Planet Labs, 50cm for Geoeye to ~10cm for that spy satellite Trump leaked recently. The MNIST dataset of hand written digits uses 28*28 pixels, but you can get away with less. Maybe 50 meter by 50 meter letters?
If you know binary you could also include extra information by using rows of 5 squares ~1m in diameter.
[0] https://support.google.com/earth/answer/6327779?hl=en
Some sort of universal symbol could be great. Especially if you knew that within X days there was a strong chance it would be spotted and rescue folks alerted.
https://nbremer.github.io/planet-globe/
I bet it wasn't.
is it? not up on the formal theory, but:
O(n) -> on the order of n x 1 time intervals of processing
O(n)^2 -> on the order of n x n time intervals of processing
etc.
so if there is a single fixed interval, N is irrelevant, leaving you with 1, denoting the single fixed interval
https://en.wikipedia.org/wiki/Big_O_notation
3D: https://www.ianitor.com/img/3dcar.png 2D: https://www.ianitor.com/img/2dcar.png
[1]I used 3789 Moon Bay Cir Wellington, FL 33414 United States
From the article ..
> *it was a neighbour who reported the sunken car and was not aware of reports that Google Maps had been used."
Yeah, I know you are just baiting me.
Probably running anomaly detection on all road-signs all over the world, you could prevent a certain number of accident each year.
The deeper question behind this, is what should have been done by those who have the info and didn't act ? Once you start running anomaly detection on data how do you navigate the ethical minefield that ensue.
Where's the minefield?
Just release it all publicly, and offer municipalities private services if they want you to really have a go at their signs.
So depending on the pricing of the API, you let accidents happen. Or even worse, depending on whether or not the municipality is rich enough to buy your services you let them happen or not (even though the poorer the city the more likely it is that they already have maintenance issues).
There are plenty other things you may notice on street-view, like dangerous/illegal behaviors, pot-holes, even finding cars in lake can have an impact of property value.
Reporting anything is probably a legal or PR risk because of the unknown ramifications so you end-up doing nothing and the accidents happens, yet no-one can hope to offer a better service because it doesn't make economical sense because of the existing service tailored to another usage of the data.
That's not an easy problem but that's one which is at the heart of all-data related business. For example, when you have access to plenty of face pictures with flash you may notice some medical issues with eyes, obviously this is a valuable information for the health insurer, and also to the user.
Actually Tesla might have a good dataset to work with since their system relies heavily on vision. Imagine location, timestamp, and sign verbiage sent to a server, if it stops seeing the sign in that area, also return an image next time a Tesla drives by for manual review, and file a maintenance request with the DOT with the click of a button.
This sounds like a job for machine learning. Are there any systematic efforts underway to use Google Earth to search for out-of-place objects?
Google is watching your backyard from above seems like a PR disaster to its already troubled image.
I think my parents had an aerial photo of their land a few years before even TerraServer was came around, from some guy who'd flown over a big chunk of territory taking photos and then sold them to the land owners. I assume this was far from the only person doing that, all over the US. Pretty cool back then, when a clear shot from the air of the area you lived in wasn't something you could grab at a computer any time you wanted, or with a cheapish drone and a few spare minutes. And before everyone else could creep on your space in the same manner on a whim, for that matter.
There is a significant discrepancy between the imagery viewable in 3D mode and the stuff that's only 2D, because Google simultaneously captures high resolution aerial photography and Lidar, particularly around cities. You can see vast differences in urban or rural landscape sometimes because the stuff you see in Google maps can be taken as far as years apart from the 3D aerial stuff.
For example, you can see where the data source and rendering method change just outside the city of Winnipeg on the South end. Probably testable in many other cities simply by seitching between google maps and google earth.
Check out how poorly the top five did in just finding large vehicles in a kaggle completion explicitly about annotating satellite images: https://www.kaggle.com/c/dstl-satellite-imagery-feature-dete...
I get where you're coming from, but there will be (probably literally) a million abandoned cars for every car worth investigating.
https://imgur.com/3jYKQ6r
https://www.huffpost.com/entry/car-with-body-of-man-missing-...
Very likely it does not have fish, but might have snakes and the occasional alligator.
https://www.waterfrontcleanup.com/blog/why-storm-water-reten...