Wildlife Detector Python Lambda Functions

I am building a wildlife detection app for a project in West Virginia. The aim of the project is to restore a parcel of abandoned mineland, and as part of that effort, we want to get an idea of the wildlife in the area. My task is to build an app that uses the ReoLink cameras already on the property and classifies the animals that they capture. This project is ongoing, and will be updated.

For the MVP, I have trained Tensorflow's built-in classification model on data from Lila BC. It is deployed to a Docker container and is invoked with an AWS lambda function. Currently, the lambda can be invoked locally via the command line, or can be triggered on upload to an S3 bucket. The resulting information is saved to a mongoDB instance.

The app features:

• Automatic image generation when a json file is added to an AWS S3 bucket
• Image generation that handles long names and special elections as well as the standard dem/republican
• Dynamically created headshot information

And was built with:

• flask
• tensorflow
• docker
• aws lambda
• mongodb