AquaHunter is a mobile web app that allows recreational lake users to learn about and report aquatic invasive species (AIS) in Minnesota. It is a citizen science app, in the sense that participants can contribute actionable data while learning about the environment. The app is intended to promote responsible AIS mitigation behavior in addition to use as a reporting tool.
To facilitate widespread use by untrained participants in the field (or at the lake), the AIS report structure is kept as simple as possible: species, location, photo, and notes. A trained professional can follow up with an additional visit to verify any potential new infestations. The precise GPS location means less time for verification, versus traditional reporting methods that only let the public report the lake or general area where they thought they saw AIS.
The streamlined interface does not require a user account. The initial form consists of a single button that captures a photo and classifies it with a neural network trained on images of native and invasive species. Google's TensorFlow library is installed on the web server (as a Python 3 celery task) as well as within the Android and iOS apps (via cordova-plugin-tensorflow). The apps provide instant classification results with or without a network connection, while the web version ensures that anyone with a web browser can take advantage of the classifier.
Once the preliminary result is known, the contributor is asked to provide a few additional metadata fields (notes, location, and optional contact information) before sending the report to a professional for review.
© 2013-2017 by S. Andrew Sheppard