A summary of the responses was requested for this thread.  Overall, this 
question received far more interest than I expected; but with the widespread 
availability of hardware producing digital images, it isn’t surprising that 
more and more people want better ways to sort through them.

There are a variety of potential software solutions:  
-Security camera software may be useful.
-EBImage
-OpenCV and Python might hold some promise, see this video for a detailed an 
amusing talk on using it to distinguish birds from squirrels.  The developer 
ideally wants to use it for bird classification. 
http://boingboing.net/2012/03/26/howto-build-a-robotic-squirrel.html
- Change vector analysis is a GIS based option that could be adapted to look 
for changes from a background photo.
-ZooImage (http://www.sciviews.org/zooimage) might be trainable for IDing in 
stationary photos. 
-Adobe products have been suggested for analyzing video, but this will still 
require manual detection of images of analysis.

 The other type of solution suggested was the crowd-sourced solution.
-Although not free, Amazon's Mechanical Turk service: 
https://www.mturk.com/mturk/welcome
could provide a useful way of paying ($0.01-0.02 per task) people to 
analyze/tag/classify photos.
- www.galaxyzoo.org was suggested as the platform where the public can sort 
through and classify images of interest. 

For image management once images have been appropriately tagged and
analyzed:
- http://gis.team.sdsc.edu/teamimages/ for a cool way (called a "faceted
search") to filter thousands of images by various properties. This was 
developed by a team at San Diego Supercomputing Center led by Ilya Zaslavsky 
([email protected]).

Thomas

Reply via email to