Skip to content
PennApps Spring 2019 Sumbission
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
images
marketing
output
videos
website
yolo-coco
.DS_Store
.gitattributes
README.md
yolo_video.py

README.md

CV-Livestock Protection

Problem Solved

Thousands of sheep, cattle, and horses are mutilated every year due to the poor means of monitoring. Unfortunately, there is no known feasible and financially viable way to get immediate notice of it and farmers get alerted only after the damage has already been done. The solution was applying Computer Vision and AI to alert farmers at a moments notice. This software detects when a predator is present in the frame and sends a text message to the farmer using the Twilio API.

Kaczensky, Petra. “Large Carnivore Depredation on Livestock in Europe.” Ursus, vol. 11, 1999, pp. 59–71. JSTOR, www.jstor.org/stable/3872986.

Who is this for?

Farmers who don't have the financial resources to invest in complex infrastructure.

Challenges & Roadmap

  • One challenge we experienced during the development of this software was creating/finding a model that accurately could identify livestock.
  • Given the limited computational capacities of recording devices, we hope to move the execution of the program on the cloud
  • Use it for monitoring the wildlife and give an ability to the researchers to get immediate notice when an event of interest to them takes place.
  • Train the model further in classifying objects that are of interest to us.
  • Develop a much more extensive front-end that shows past information/threats along with adding location input.

Technologies

  • Twilio API - REST for Text Messages, with custom number.

  • YOLOv3 - Real Time Object Detection

  • TensorFlow - An open source ML framework for everyone

You can’t perform that action at this time.