I don't think this is satellite data, most likely photos captured from planes.
And I'd kinda be curious to what Microsoft has been working on, I remember their birds eye was pretty good a few years ago.
Apple I'd believe the are selectively trying to improve their data in areas that are heavily visited by their users.
Also, I did for the first time see an Apple branded car with a photo sphere on the roof recently and kinda got excited that they are gonna try and bring competition into maps / street view that Google has dominated for the last decade.
Essentially they use a rig of DSLR cameras mounted to the bottom of a Cessna to take photographs in a 45 meter grid, and then do stereo reconstruction from that. Crucially they describe using fast phase based disparity maps, which at the time were a recently declassified technique for guiding cruise missiles. C3 was a part of the defense contractor Saab AB.
Saab also owns Vricon, basically the same thing as C3 but not aimed at consumers. They sold part of it to DigitalGlobe, and they use both aerial and satellite imagery for 3D models.
Nokia also invested in C3 and had the first 3D maps available to consumers before anyone, completely in WebGL which used to be on https://wego.here.com. I guess Apple had more money to acquire them.
Considering the expense, and especially time, involved in gathering street-level data, if I was Apple, I'd put every sensor I could think of on those cars and then let the data scientists figure out a use for it later.
Colossal bandwidth use for very high latency updates? Surely it’s better to get depth information from the device rather that from Apple’s self driving cars/StreetView clones, either with stereo, lidar, or AI estimates?
It'll be planes, because the running cost per mile travelled is critical to making the whole thing cost effective.
Power consumption and weight of the rig isn't insubstantial, so the smallest planes and gliders can't be used, and after that cost per hour and cost per mile are the key metrics.
>I did for the first time see an Apple branded car with a photo sphere on the roof recently and kinda got excited that they are gonna try and bring competition into maps / street view that Google has dominated for the last decade.
I recently asked the founder of a pretty big GIS company how they handle plagiarism. The answer: "we don't bother. [we just monetize services on top of open data.]"
So it looks like the intentional errors method didn't work out in the end. [1]
Wonder if Apple will be as accommodating though :p
Coincidence: Half an hour after reading this thread, a white Subaru with California tags stopped at the red light outside my office window. Big disco ball on a platform attached to the roof and "Apple Maps maps.apple.com" stenciled on the window.
Mostly used popular tools like Hopper, LLDB, Hex Fiend. Xcode's debugger UI was surprisingly helpful too because it lists current memory allocations of obj-c instances nicely. Also learned about Frida along the way and would have used it more in retrospect.
And I'd kinda be curious to what Microsoft has been working on, I remember their birds eye was pretty good a few years ago.
Apple I'd believe the are selectively trying to improve their data in areas that are heavily visited by their users.
Also, I did for the first time see an Apple branded car with a photo sphere on the roof recently and kinda got excited that they are gonna try and bring competition into maps / street view that Google has dominated for the last decade.