Inspiration
We want to assist organic farmers in increasing their yearly yields and by giving them a tool to quickly identify bad crops so they can dispose of them. By identifying a crop with a disease, the farmer has an understanding of the state of the soil and the pathogens her/his crops are exposed to. This also allows him to find crops that are resistant to pathogens, which she/he can cross-fertilize to genetically improve the immunity of future crops.
What it does
Uses video feed to identify whether a plant has a disease or not in real time.
How we built it
Trained a convolutional neural network and deployed it into an android app with Android Studio.
Challenges we ran into
- Long training periods because number of parameters was too large - we had to fiddle with the model.
- Large amount of time spent on data pre-processing ## Accomplishments that we're proud of
- Real-time detection
- 92% accuracy on training, 80% accuracy on validation set ## What we learned
- Convultional Neural Networks
- Keras and Tensorflow
What's next for artificial intelliplants
- Improving the model
- Local weather fetching to obtain farming season statistics
Log in or sign up for Devpost to join the conversation.