Green Stems

Assisting in the garden health and identification, just with a click.

Comment

Inspiration

As we know maintaining a healthy garden is tough and needs a lot of care and nourishing to have good and healthy growth. Sometimes the early sign of any disease or anomaly in health is left unobserved, which leads to the death of those saplings. There is a need for early detection of any disease and medication. Also with the health circle, sometimes the identification of those saplings remains undone, leading to a lack of right remediation. So, there was a need for an overall app that can take care of all these needs.

What it does

  • It is an application built with AI capabilities for disease detection, species identification, and providing remedies for those diseases and issues.
  • It also has a page dedicated to gardening articles and blogs for the overall care of the plants and greenery that can help newbie gardeners kick-starting their gardening journey.

How we built it

Used Android Studio and GPU-supported cloud platform to process the humongous data and deploy the models.

Challenges we ran into

  • We never worked on the deployment of a model to an app before.
  • We never worked on image processing with large datasets before.
  • Some datasets were not suitable, so we worked on the pre-processing to make it suitable.
  • Never worked on the cloud for model preparation before, ran into some issues but we fixed them.

Accomplishments that we're proud of

  • We worked on a plant-focused application for the first time, researched over the same, and tried to build something which can help anyone who loves plants and have a garden.
  • For half of the team members, it's their first hackathon and they did a great job in building something so quick.

What we learned

  • Deployment & scaling application using Android Studio
  • Image processing of huge data
  • Pre-processing of an image-based dataset.

What's next for Green Stems

  • The dataset is still not as accurate as it should be to predict every possible plant species. We will improve the dataset used to train the model to get the best results for all possible local plants found in the garden.
  • The user interface will be tuned to make it more user-friendly with calming UX and UI design.
  • We're also planning to add a feature to connect with local peers for gardening sessions and meetups to build local communities to spread awareness.

Built With

Share this project:

Updates