Crop Disease Detection

An android app + NN backend to detect crops from images

The project received the Second Runners Up a Microsoft Code.Fun.Do hackathon, IIT Bombay.

This application uses a modified AlexNet architecture to classify among 38 classes of healthy and diseased crops. We use the PlantVillage dataset for training the network.

The app also detects chances of a plant epidemic by compiling the results of diseased crops over certain geographical regions.

Github source code is here.