Case study
Puccinia Polysora Detector is a tool that help you check if your corn is infected by the puccinia polysora disease It features a two layout grid, the left is an historic and the the right contains the main app feature where you upload a picture of your leaf and it predict it to you
Role
Full-Stack Engineer
Scope
Full-stack ownership
Stack
Python, Streamlit, SQLite3, ScikitLearn
Problem
Food security in Madagascar depends heavily on maize cultivation. However, maize crops are regularly threatened by fungal diseases, particularly Polysora Rust (Puccinia polysora), which is especially devastating in the island's warm and humid regions.
Solution
Puccinia Polysora Detector is a tool that help you check if your corn is infected by the puccinia polysora disease It features a two layout grid, the left is an historic and the the right contains the main app feature where you upload a picture of your leaf and it predict it to you
Product
1 / 5
Ownership
Searching data accross the web for puccinia polysora detection (rare data)
Processing data by separating hue of each pictures to isolate the yellow mask (puccinia trace)
Separating datasets into three: training ,testing, validation
Choosing the best decision tree and random forest from Scikit-learn by doing some benchmark and perfomance test
Teach the model
Build the streamlit app to work with the model
Stack
Python
Streamlit
SQLite3
ScikitLearn