3 ms·
I'm not sure you appreciate the engineering infrastructure at work here. This isn't special because Google is doing it, it's special because it's really hard to
by dylangs1030 13y ago
I'm not sure you appreciate the engineering infrastructure at work here. This isn't special because Google is doing it, it's special because it's really hard to do. How would you go about attempting this project as effectively as Google with the resources of a typical startup company?
- toomuchtodo 13y agoCrawl the web for images, use the GPS EXIF data to find clusters, and use image processing to stitch together 3D environments from the photos. It would not be as good as streetview though.
- dylangs1030 13y agoCreative...but I don't think it would have the same fluidity to Google's Trek. The images wouldn't be as up to date or "real time" as they could be because you'd be caching images that are potentially several days or weeks old. And if you scripted a time limit for images, e.g. only within a few hours, you'd heavily limit the amount of usable images you had. With "agents in the field" Google could potentially update its feeds whenever and as often as it needed. The difference would be having an in-house image-generating infrastructure or...well...tag-searching Instagram.
- mh- 13y agoMicrosoft did this, effectively. Someone who remembers the project can provide a link to the demo/talks (TED, if I recall) hopefully. edit: believe it was related to http://en.wikipedia.org/wiki/Photosynth http://en.wikipedia.org/wiki/Photosynth
- toomuchtodo 13y agoEDIT: Looks like I'm behind the times. Google already did this: http://www.fastcodesign.com/1669688/googles-photo-tours-suggest-a-crowdsourced-3-d-world-map http://www.fastcodesign.com/1669688/googles-photo-tours-sugg... http://google-latlong.blogspot.com/2012/04/visit-global-landmarks-with-photo-tours.html http://google-latlong.blogspot.com/2012/04/visit-global-land...