Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

360 "birdseye" parking assist

This site may earn commission on affiliate links.
This topic comes up a lot. I know that Tesla does NOT have the hardware to make this happen in a conventional stitch and process the image sort of way, but......

Since we know that the FSD/AP computer is constantly generating a 3D map of our surroundings (think = cones, trash cans, stop signs), couldn't Tesla take a technically creative approach to offer us the approximate 360 view? For example, Tesla could stitch together a 360 degree image of what the cameras currently "see". This would leave a lot of voids closer to the car where the cameras just don't cover. BUT, the camera HAVE probably already seen the curbs or cement barriers or ?? as you pulled into the spot your car is currently in, so couldn't they just render the 3D map that the computer has into the missing voids as an approximation? I don't know how useful this would or wouldn't be. But, it seems like an interesting technical challenge that might yield results and I'd be curious if there is a solution here.

Edit: Oh, and if you're one of the "good drivers don't need this $h!t !!" comment-types, I'd invite you to move along....nothing to see here ;)
 
Since we know that the FSD/AP computer is constantly generating a 3D map of our surroundings

Not only to do we not know that, we know it's NOT doing that.

They're currently re-writing it to do some version of that, but it's not here yet.


couldn't Tesla take a technically creative approach to offer us the approximate 360 view? For example, Tesla could stitch together a 360 degree image of what the cameras currently "see". This would leave a lot of voids closer to the car where the cameras just don't cover. BUT, the camera HAVE probably already seen the curbs or cement barriers or ?? as you pulled into the spot your car is currently in, so couldn't they just render the 3D map that the computer has into the missing voids as an approximation?

Once the re-write is done, yes it'd theoretically be possible they could program it with some idea of object permanence where it shows what it thinks WAS in the spots it can't actively see.

It'll still have some blind spots it never saw when making a tight turn around an object or something...or things that moved into blind spots when the car was asleep like a small animal or a basketball or something (or things that moved out of em it though were there before it went to sleep)... but it'd still be better than what we have now where, say, a curb (or high bumper on a truck) simply disappear to the car when you get close enough to them.


I'd honestly expect this to almost be a requirement for something like reverse summon to work (and also for it to finally start using cameras, not just ultrasonics, for self- parking).
 
  • Like
Reactions: beachmiles