When you think about it, Google Street View is a rather logical offspring for Google. What made Google the successful company it is is its search engine for the web, right ? And a web search engine is basically a tool to explore the web, some sort of compass in the virtual world. And Google Street View is a tool to explore the real world.
From that standpoint a logical extension for Google Street View would be to be able to search the world using keywords to look-up the GSV database. For example if you are looking for “Barbershops in Philadelphia” you would get street views of all identified barbershops in Philadelphia, it is that simple. You may say that it does not seem different than looking up for “Barbershop Philadelphia” with Google search engine and then click on the map from the search results. But actually it is, because the idea would be to search for visual artifacts captured by GSV cameras – as opposed to searching web contents linked to a physical address.
A more appropriate example is searching something like “green house Philadelphia” that would return every green houses captured by the GSV cameras. Actually the idea here is to make an enhanced version of GSV that is the exact equivalent of what a search engine is for the web.
Now of course that would require a technology which is able to analyze the visual content of GSV and arrange it into a database. Though I am no expert I guess such thing is not a technical utopia. Ideally the content should be accessible in its most objective form, i.e. not be filtered. That means that not only the permanent features will be searchable (“green house”) but also the artifacts and accidents captured. For example you should be able to search for “cat” or “bike” or “kissing couple” as well as “red house”. That would make the system more open, suitable for seemingly useless purposes – and thus more interesting.
Another limit of the existing system is the quick obsolescence of the data captured. A web search engine crawls the web incessantly to update its database and index. A “real world search engine” needs to be able to do the same which means that the GSV “camerabots” would need to crawl the streets almost everyday. That would add much more depth to the system and makes it suitable for important documentary survey and to answer critical questions such as “Were green houses in Philadelphia more numerous than red ones in 2015 ?” or “In which city is it the most likely to stumble upon a cat when crossing the street ?”. Now you may think that updating the Street View image database on a daily basis is impossible, but you are wrong.
Actually a rather simple way to achieve that is for the operator to involve people like you and me in the process instead of having its own fleet of cars with mounted camera. Would you volunteer for having a camera installed on your vehicle and you’ll get some monthly fee and more importantly your name will be credited on-line. Taxi companies could also benefit financially from it in the form of a partnership.
I am personally neither supportive nor opposed to such a tool. It is merely a speculation of what is arguably a plain logical extension of what GSV is with its current limitations. Therefore it would not be surprising if it develops in such a way.