Rice Plant Disease Detection

Project Domain / Category

Image Processing / Artificial Intelligence / Web App

Abstract / Introduction

Rice is one of the topmost agricultural products. A billion people around the world rely on rice as their main food, making it an important crop in most countries that grow it. However, diseases such as (Leaf smut, Brown spot, Bacterial leaf blight) that affect rice leaves can be a major problem because they slow down plant growth and can ruin the crops. Identifying leaf diseases is important for sustainable rice farming. Manual identification of diseases is expensive and inefficient, highlighting the need for efficient detection methods. Traditional ways of identifying leaf diseases in crops, like rice, are often slow and not very effective. This situation demands an automated computer-aided diagnosis tools and techniques to detect diseases. The proposed project aims at performing multi-class classification of classifying rice plant images into three classes (Leaf smut, Brown spot, Bacterial leaf blight). You are required to develop a web app in which the user will enter images and check the status.

Functional Requirements:

·         Dataset Collection: Collect image datasets from available free repositories or any other online source.

·         Pre-Processing: Use different image processing techniques to create a uniform, normalized image dataset. You may need to perform data augmentation in this step.

·         Model Selection: Analyze different deep learning-based CNN models and select a suitable one for classification.

·         Dataset Splitting: Split the dataset into training and testing set for model evaluation.

·         Model Training: Train the selected model using training dataset.

·         Validation and Hyperparameter Tuning: Validate the model's performance using the validation set and fine-tune hyperparameters like learning rate, batch size, and network architecture to achieve the best results.

·         Model Evaluation: Check the performance of the model used using testing dataset and different evaluation metrics.

·         Real-time Detection: Implement a real-time rice plant disease detection pipeline using OpenCV to upload an image from and apply the trained model for rice plant disease detection.

·         User-Interface: Develop a user-friendly interface in which the user can upload images for analysis. The interface should provide visual feedback, such as original images alongside classification results.

Prerequisites:

·         Have a good understanding of Python.

·         Having knowledge of basic deep learning concepts and models.

·         Understanding of basic image processing techniques (preferable but not mandatory).

·         Basic idea of working with image related datasets.

Tools:

·         Language: Only Python

·         IDE: JupyterNotebook, Pycharm, Spyder, Visual Studio Code, etc. Better to use Google colab environment or google cloud.

·         OpenCV

Note:

·         VU will not provide any kind of paid resources needed for the project.

·         A student must find the dataset by himself / herself.

·         Use of any other language is strictly prohibited.

·         Kindly read the instructions given properly and choose a project only if you have developed a clear understanding of the project.

·         A student who wished to select this project must commit to spend 2 hours daily for FYP project. This may include learning through tutorials or getting help from any reading material.

·         In case of any query, feel free to contact and discuss with me.

Important links and Tutorials:

·         Python

·         Deep Learning

·         https://www.w3schools.com/python/

·         https://www.tutorialspoint.com/python/index.htm

·         https://www.programiz.com/python-programming

·         https://www.simplilearn.com/tutorials/deep-learning-tutorial/guide- to-building-powerful-keras-image-classification-models

·         https://www.analyticsvidhya.com/blog/2020/02/learn-image- classification-cnn-convolutional-neural-networks-3-datasets/