Augmented reality application have developed rapidly over the last year and have reached a number of platforms by now. Also there are now multiple providers to run AR services. This is a dramatic change from the early days were Layar was the only open plattform. Same thing with location based services. Google Latitude and Brightkite have been overtaken in the meantime by Foursquare and Gowalla as the leading applications. Actually one dosen't even think of them as applications but communities.
Anyway to push their position Google now tries a new approach, integrating image recognition in to the AR environment. This could potentially improve the service in terms of accuracy. So far the technology relied on the location from the aGPS and the direction via the compass. However this is within a range of a couple of meters up to ten twenty meters. So in this sense you could have actually already walked past the trendy Sushi place you are desperately trying to find.
Image taken from AddSite / Scanning a contact card.
However, now with Google Goggles (its sounds a bit like Googoo Goggles, the Dr Seuss character) your mobile client will detect the place via its features, scanned from the camera image. It still links back to a massive database containing the background information but the identifiers were delivered by the camera. In this sense a true visual search.
It is still in beta and overall the technology is still in the beginnings. Unfortunately it runs on Androids only. Currently it can scan contact cards and translate them into a digital contact on your phone, recognise art work (I would like to see it recognise this Giacometti sculpture?), recognise landmarks (incase you are usure whether or not this is the Eiffel Tower you are looking at), detect logos (this could be helpful with the sushi place) or also book covers and presumably posters like theater plays or movies.
Search and related services are the core business of Google and this is the sort of innovation they are looking for. Already the term visual search makes a lasting impression, and linkes to all sorts of relations.
It sounds like a great application of technology and it will probably work very well at some point. However on the other hand it does start to rais the question of how dependent on the mobile client do you want to be? Do you really want to look at the world through the ridiculous small screen of your touch screen? Maybe to fins the sushi place?