erengulum/plantcare

predicts for leaf type, infected or not and location of leafs

PlantCare

• Web application for leafs: https://bbm416-plant.herokuapp.com/

Specifications:

• Predicts leaf type, whether infected or not and location of leafs in 10 seconds.
• Recognizes apple, blueberry, cherry, corn, grape, orange, peach, pepper bell, potato, raspberry, soybean, squash, strawberry and tomato.

Dependencies:

• PyTorch
• YoloV5
  • modified, some imports are deleted to fit to Heroku.
• OpenCV
• Django
• Pillow

Run:

• Clone the repository, run pip install -r requirements.txt, python manage.py migrate, python manage.py makemigrations then python manage.py runserver

Datasets:

• PlantVillage
• PlantDoc

Images:

home page of the site

prediction of given image

Bugs and Limitations:

• Due to Heroku, CPU version of PyTorch is used, so it is little slow.