I know this stack exchange provides the ability to post photographs of animals/plants for species identification, but is there a web resource that can automatically identify a plant/animal using image processing/correlation tools - so the user doesn't have to painstakingly work with keys?
$\begingroup$
$\endgroup$
4
-
$\begingroup$ Try searching "species identification website" and there are many websites for species identification. Therefore I have flagged as too broad. Also, you haven't showed much research effort. A simple search gives you so many results. $\endgroup$– TanMathCommented Nov 5, 2015 at 2:26
-
1$\begingroup$ I recommend using iNaturalist and their mobile app. That is very useful community in species identifications. $\endgroup$– DexterCommented Nov 5, 2015 at 5:19
-
2$\begingroup$ Encyclopedia of Life is also a good database for species identification but it does not have "automated tools" to do the search. Any kind of image search would work reasonably well if the training set is big+diverse enough or the query is unambiguous. This is a machine learning problem. Yes the user has to take pains and that is why there are biology courses on these topics. $\endgroup$– WYSIWYGCommented Nov 5, 2015 at 6:20
-
1$\begingroup$ Quality of images, variety in organism physical traits, and institutional motivation & coordination, or lack thereof, are the primary reasons why a tool like this doesn't exist. To automate species IDs, extremely specific images would need to be provided; e.g., high image resolution & quality (TIFF files), region(s) of organism depicted (ex. reproductive organs), age of organism, size/distance away from organism, etc. All of these parameters would need to be standardized, and then strictly followed. $\endgroup$– user22020Commented Oct 18, 2017 at 15:11
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
Scientists are only beginning to develop robust techniques using DNA markers and NIR spectroscopy, for instance. For automated systems based on photographs there are some tests too, but usually restricted to very narrow taxonomic groups.
