tnc-ca-geo / animl-frontend

A frontend web app for viewing & labeling camera trap data by The Nature Conservancy.
https://animl.camera
Other
18 stars 0 forks source link

Allow management of field data about deployments #207

Open abfleishman opened 8 months ago

abfleishman commented 8 months ago

Many camera trap management software options include the management of all of the site/deployment metadata that is collected by typical camera trap practitioners. The aids in downstream data management and facilitates data export with all the info that a user might need for the analysis attached to the annotations.

This probably is not something typical of invasion detection workflows because the only important info may be datetime, location, and species for that objective. However, picture a meta analysis a few years into a project when you want to answer questions like:

  1. "Does camera height above ground affect cat/rat detection rates?"
  2. "Does camera view direction (n/s/e/w) affect trigger rate during daylight?"
  3. "For invasion detection of rats is it more effective to deploy cameras on trails or randomly?"

There are loads of "custom" data fields different researches might use about their delpoyments/locations but there are probably a bunch of common field that people use all the time and are used in most studies. It would be interesting to consider how to add additonal data into the model for animl.

A resourse on what people often collect may be the Remote Camera Survey Guidelines & AB Metadata Standards as they seem to have done a pretty thurough job compiling info for surveys in western canada. I bet a look at a few of the other platforms out there (aguti, WI, trapper, TagTrapper) would also be very informative.

postfalk commented 3 days ago

Do we need fields here. Can we just support a JSON object. In the user interface it would look like, add field and then a value. It would be on the user to keep track of their fields.

abfleishman commented 2 days ago

I don't see why not, but You need to be able to associate the JSON with specific images based on metadata (datetime + sensor_serial).