The AI Image Toolkit, or AIT, is designed to manage camera trap projects including the processing of all camera trap images collected in the field.  Biologist can "tag" the observation records with various attributes found in the image, such as the species, gender, age class, behavior, and more.  Various data exports and reports are available to summarize animal presence and activity at a location.

This tool uses Microsoft's MegaDetector v5, an artificial neural network which has been trained to identify animals within images.  When images come in from the field, they are processed by this library and those images which do not contain an animal are discarded.

This tool makes it easy to export data from individual sites.  Additional exports will be made available by species and other attributes which will span locations and projects.

This tool provides a tagging style editing ability across projects based on the user's roles assigned.  Users with the "data manager" role can create project and locations and overall structure of the project while "taggers" can add observations (upload images) and update their attributes once they have been run through MegaDetector.

Acknowledgments

Thanks to the Microsoft MegaDetector team.  Its all about helping Wildlife, and they get it.

The animal list was initially populated with data from the California Roadkill Observation System (CROS).

We would like to thank the contributors of the open sources tools and technologies which we used in this project, including:

• Drupal Content Management System, from the core contributors to third party module developers
• W3-based website theme developed by Alaa Haddad at (https://www.alaahaddad.com/ and https://www.drupal.org/project/d8w3css)
• Server based technologies which this tool relies on include: drush, composer, exiftool, rclone
• Cloud hosting Services: Microsoft Azure