Inspiration
It is really difficult to detect diseases of plants, and even more when you are new to this sector. I have seen many people have no idea how their plant is doing or which disease they have. Many villagers need to send their plants to labs to check out their diseases which take more than 2-3 days.
What it does
This application takes the image of diseased leaf/fruit of the plant and gives the exact disease name. It works in real-time as well as we can upload it from the gallery. This helps to ensure we can work on any device.
How I built it
I used the android studio as the base of the application. Using TensorFlow to create a CNN model with the use of Plant Village Dataset and made the model into a TensorFlow Lite model to use it in android. Just the basic UI for android with 3 buttons.
Challenges I ran into
There were many difficulties converting the TensorFlow model into the TensorFlow lite model as it was my first time doing so and it took a lot of research.
Accomplishments that I'm proud of
The confidence of the model is very high and I am very proud of the results of the diseased plants.
What I learned
I learned how to change a tensorflow model to tensorflow lite. How to make a CNN model from start to finish and also how to make an idea into reality.
What's next for Plant Disease Detection App
Remedies for the plant diseases, IoT interaction, and even realtime e-commerce is the future for this app. An ever-expanding project.
Built With
- android-studio
- java
- keras
- kotlin
- tensorflow
- tensorflowlite
Log in or sign up for Devpost to join the conversation.