Gotta Catch them all! (The Pests)

Based on the recent success in computer vision, we propose to fight the problem of pest detection with algorithms that can easily run on the edge devices & smartphones.

Comment

Inspiration

Agriculture industries contributed $25.1 billion or 51% of GDP in Canada and this market is expected to grow up to $85 billion by 2025. This project will provide a small incremental benefit of doing pest detection on low resource device to the end users.

Technical Insipiration

  • Tensorflow has recently released TFLite, a library that allows developers to build on-device models that provides personalized, low-latency and high-quality recommendations, while preserving users’ privacy.

What it does

Identify pests, insects efficiently with EfficientNet, an optimized, low-latency solution that can be used on low resources devices like RaspberryPi, Smartphones.

How I built it

  • Tensorflow TF Lite library
  • Jupyter notebooks & Google Colab
  • Web scrapping
  • HTML, CSS, JS for the looks

Challenges I ran into

  • Video Production :-)
  • Curating dataset

What I learned

  • DMADV Methodology by Dr. Amy. Planning to use it more frequently.
  • learned a lot about the different problems that arise at different stages in agriculture.
  • Got introduced to Carbon sequestration, soil erosion.

What's next for Gotta Catch them all! (The Pests)

  • Integrate with
    • Android devices
    • Intel openVino platform

Built With

Share this project:

Updates