- In real world, farmers face lots of devastating loss only due to they don't know which disease is affecting their crop. This project is mainly focused to solve that problem.
- Here I take images of corn, potato and tomato which is affected by the disease Corn-Common_rust, Potato-Early_blight and Tomato-Bacterial_spot and train it on custom CNN model.
Our pipeline consists of three steps:
- An AI model which detect plant disease.
- An AI model which predict if the leaves has disease or not.
- The output is predicted disease name.
- Our Custom CNN model perform better by giving near 95% accuracy.
-
Clone the repository:
https://github.com/dipesg/Plant-Disease-Detection.git
-
Create separate conda environment:
conda create -n plant python=3.6 -y
-
Activate environment:
conda activate plant
-
Install all the requirements:
pip install -r requirements.txt
-
Run following script to run the program:
streamlit run app.py
- Pandas
- Numpy
- Matplotlib
- Sklearn
- Tensorflow
- Streamlit
- OpenCV