h a l f b a k e r yVeni, vedi, fish velocipede
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
Please log in.
Before you can vote, you need to register.
Please log in or create an account.
|
Crowd Mapping
app that lets phone-users integrate their personal street-view into coherent information | |
I suggest putting all the camera phones with GPS to good use - let users upload pictures from their phones to a site that pays in whatever currency optimises the real worth/felt worth relation (ringtones? status-points?).
The pictures are used in a variety of ways: As the phone-type is known, the
exact properties of the sensor are known : From a picture of Big Ben(identified via image recognition) + the current location of the user various things could be computed: smog/fog levels and sunnyness, for instance. From a picture of a highway, the current traffic could be extracted, from a picture of a 'closed' sign the status of a highway/ public toilet/marketplace.
All this hinges on the exact location of the user, the sensor (specific camera parameters), and some direction (possibly extracted from visible landmarks) being known. To optimise the process, users should get boons for extracted information, not for any shot they upload, thus motivating them to make better shots. Of course the users could also input the relevant information by hand, but i guess just uploading a picture via an app is easier than trying to convey just how much fog there is, or which lane of the highway is closed.
The database finances itself by providing real-time map updates to companies.
The database could also put out requests for specific angle-shots, that the app on the phone sifts for applicable locations/camera requirements - ' Wide-angle, Low noise shot of waterfront from SSW at 94948/29384 needed in next 5 min. 5 ringtones free.'
You, in somebody else's vacation photos
http://web.archive....20vacation_20photos
[hippo, May 28 2009]
GPS_20Camera_2fvideo
...going back to 2000... [hippo, May 28 2009]
Pictures_20-_20The_...they_20should_20be!
...a not dissimilar idea... [hippo, May 28 2009]
Sightseeing_20Experience_20Sharing
[hippo, May 28 2009]
[link]
|
| |
It's a good idea, however the ability to implement such a tihng is difficult still since computers are consistently horrible at identifying useful things in an environment. What would probably work better would be a team of real-life human-beings who sit and wait for new images to come in to help identify what is in the images for some compensation (the more people vote it up from the people who collectively answer it the higher their chance of being compensated for their efforts). In essence you'd have high competition in identifying things in images (almost immediate results). |
|
| |
[quantass, please send me email at bakesperson@halfbakery.com. Thanks! --jutta] |
|
| |
[21_Quest]: Might be google dished out the wrong destinations on my search for the apps you mentioned, but they are not even near to what i proposed here. World Traffic Cams seems to at least give information about a highway, but from webcams, and by showing you the pic of the highway. That is far from an app that takes users' photos, analyses salient features, and gives out information about them. |
|
| |
Panoramio is a nice app, but it is just a picture repository. |
|
| |
traffic patterns aren't possible from single pics, true. But i posit that it already is possible to give a rough estimate of cars visible on a stretch of highway. Numer of cars per highway gives rough estimate about the traffic, while more than one pic, over time, gives estimate about pattern. |
|
| |
If there is a pic of big ben, at a known time of day, from a known angle, by a knwon sensor, i posit that it is possible to say something about visibility at the time of the next picture with all those variables known, if there are other pics indicating the cloudiness at that time etc. |
|
| |
But positing is just another word for dreaming, 's true. |
|
| |
I am not suggesting that this would work from day one, or perfect, but as with the case of rain detection by cellular-network signal-goodness patterns, it is better than nothing. |
|
| |
I have read of a piece of software that trawls Flickr for images tagged with the name of a certain landmark or building, and reconstructs a 3D model of said building from the 2D images. From memory, the published test used Notre Dame de Paris, and came up with a startlingly accurate model. |
|
| |
If one were to search by GPS tags on uploaded photos, presumably this same software (or an adaption of it) could, over time, build up a 3D model of the entire inhabited world. |
|
| |