Plant Leaf Disease Detection

In our research, we have used Convolution Neural Networks (CNN) to detect and identify the type of leaf as well as the disease it has been affected with. The image dataset we used for the training purpose is titled ‘PlantVillage’. In this the plant species were properly differentiated with respect to their species and disease they have been affected with. This image data was first made compatible with our CNN model by reducing its dimension to 227X227 pixels. The model was trained on various CNN layers to ensure that the features are extracted suitably. The base technology we have used to prepare the model is TensorFlow which is been developed by Google to carry out various deep learning functionalities. The project is been divided into 6 different modules Accusation of Image, Preprocessing of the Image, Feature Extraction, Disease Comparison, Disease Detection and Final Output. To deploy the model we used streamlit web application python module. Here we could upload the leaf image and for our model to make a prediction. With our model we were able to achieve an accuracy of above 98%.