Worldwide Street View in a week ?

[oldtobor]

I was using the street view feature of maps.google.com. You can actually see the streets of some major cities and some suburbs like Philadelphia or Miami, etc. I noticed that you can't view any place outside the USA. If everyone in the world sent a youtube video of the view of their local streets to youtube, in a week the entire world could be covered! You could see how every town and city on Earth is like!

I was just thinking of how easy it is to obtain such a massive amount of information in such little time by just letting everyone upload their own videos to youtube. I imagine that the concept of Web 2.0 is much more powerful than most of us realize. How much do you want to bet that all of a sudden on youtube you will magically find all the street views of every possible place on Earth ? Hmm...

But wouldn't that actually be bad ? You destroy the mystery of the world...

 

[brewnog]

I wonder why you can't see anything outside the US; every address I've ever lived at in the UK is covered.

The information is already on Google Earth. I don't understand why you would want to upload it to Youtube.

 

[oldtobor]

The Street View feature is only available from maps.google.com as far as I know. This shows you a clear picture of the street and homes from the sides and all around. I am not sure if you can see it from google earth, but anyways it is only limited to some US cities, suburbs.

If everyone in all parts of the world took a small video of their street and area and uploaded it to you tube, then you could enter address, city and view almost any place on earth. Since it is a parallel process, you could achieve full coverage in a very short time.

I am not sure if street views will ever be able to cover all the world being that it is a serial process. One company doing it against thousands of individuals doing it.

 

[mgb_phys]

Imagine the work needed to deal with thousands of videos in different formats with messages saying roughly where they were.
Then trying to calculate projections, orientaitions and distortions from low quality mobile phone video and fit different image sources together to try and work out where they were facing.
It's much easier to drive a multi-camera system with GPS and interial sensors to say where you are and have regularly spaced high quality images in known directions.

 

[oldtobor]

Imagine the work needed to deal with thousands of videos in different formats with messages saying roughly where they were.
Then trying to calculate projections, orientaitions and distortions from low quality mobile phone video and fit different image sources together to try and work out where they were facing.
It's much easier to drive a multi-camera system with GPS and interial sensors to say where you are and have regularly spaced high quality images in known directions.

True, but the you tube or google site or whatever would filter out the bad videos and upload only the good ones. And then there could be some economic incentive to those doing it, or google could rent out the cars worldwide and let people do it. The point is, it could be done very quickly and worldwide. That would greatly change our perception of the world, it would really mean alot. If the images were high resolution, there is no limit to how much of the world you could see. It is a very interesting and intriguing thing. I don't know what the philosophical consequences could be ...

 

[mgb_phys]

It's a much harder job than you imagine!
First you need high accuracy GPS to give you a >0.5m position, you probably also want intertial sensors on the platform to handle the car's motion. The cameras need to be calibrated and accurately positioned relative ot each other with overlapping fields of view.
The system used is this one - http://www.immersivemedia.com/products/products.php?pageID=70
The result isn't just a movie out of a car window - with this system they can build an accurate 3D CAD model of the cities they drive through.

You could do it with just ordinary video if the blogger had also taken a theodolite/total station and measured the x,y,z coordinates of about a dozen features in visible in each frame!