This code repository contains Dockerfiles and Python code for camera trap animal detection and classification. Our container based approach to camera trap image processing is modular, easily scalable, fast to deploy and lowers system prerequisites. Developed by RZA, JCB and BWB.
The /examples directory contains command line scripts that can be run on an appropriate system. The scripts pull a Docker iamge from Docker Hub and then perform the processing on your camera trap image folder. For GPU support you need to use a current version of Windows 10 or Linux.