Google Blogoscoped has raised an interesting idea: could Google use OCR on their StreetView imagery to improve the accuracy and/or functionality of Google Maps? There are a number of intriguing possibilities:
- Simply make the maps more accurate. If you can read an address off of a building, or the name of a business, you’d know exactly where to place the pin in Google Maps.
- Make text searchable. Even if a new location wasn’t in Google Maps yet, you could still help people find it if they searched for the right keywords.
- Overlay translations. If you are browsing a location that isn’t in your native language, Google could overlay translations of the words found in the photos.
- If Google could update the imagery more often, you could find sales, job openings and other tidbits like that. I can’t imagine they could update it that quickly, but you never know what might happen a few years down the road.
Other suggestions from the Blogoscoped forums include inserting ads (maybe even overwriting billboards), or using face-detection to try to identify criminals. I think Google would face some serious opposition to the second idea (and maybe the first), but it’s interesting to see the ideas that people come up with.
The tricky part to this is obviously the OCR. You’ve got relatively low-quality photos (at least that’s what they show us), taken with a wide variety of light, at different angles, different fonts, etc. It wouldn’t be easy. However, applications like Evernote with its amazing OCR can give us some hope. Google certainly has the technology and the horsepower to get this done, but I wonder if they’ll try it or not?
What other cool things could Google do with this if they decided to implement it?