AnimalSight | Devpost

Inspiration

There are lots of animals I come across in my backyard but I often don't know what they are. We think urban animals are underappreciated and not enough people know or care about them which is why we decided to make our project.

What it does

Identifies animals in real time from a camera feed and gives you more information about them including how endangered they are, if they are dangerous to humans, and other fun facts using ChatGPT API. The homeowner and animal protection agencies are automatically instantly messaged / called about information about the animal location and sighting time when a highly dangerous or endangered animal is spotted. If there is a certain animal you have been waiting to spot in your backyard, this can also be customized to message when that particular animal is spotted. It can also be used by farmers to track and identify when hostile animals that harm crops are spotted. The primary purpose of the project is as follows: 1) Increase public awareness and appreciation for backyard animals 2) Alert wildlife conservation authorities when dangerous / endangered animals wander into urban areas 3) Data collection and visualization for urban wildlife patterns

How we built it

1) OpenCV for computer vision and YOLO, Google Collab, Roboflow for the animal detection algorithm (custom CNN) 2) Python + Flask for backend 3) HTML, CSS, JS for frontend 4) APIs: - Twillio for automated SMS/phone calls - ipinfo.io for getting realtime gps coordinates for GPS animal heatmap + messaging info to conservation authorities -ChatGPT for information about animals 4) MongoDB for database; Matplotlib for data visualization and graphs.