Leaf Disease Detection Using Computer Vision

Introduction

In the realm of agriculture, early detection of leaf diseases is crucial for maintaining crop health and ensuring high yields. Leveraging computer vision and deep learning, I developed a project to detect various leaf diseases. This article provides a detailed explanation of the project, including the code and workflow.

Project Overview

The project involves training a Convolutional Neural Network (CNN) to classify images of leaves into different disease categories. The workflow includes data preprocessing, model training, and creating a graphical user interface (GUI) for testing the model.

1. Data Collection and Preprocessing

The dataset consists of images of leaves with different diseases. The images are organized into training and validation sets.

Code Explanation

• Imports: Import necessary libraries from TensorFlow and Keras for building and training the CNN model.

Model Architecture

• Model Layers: Conv2D: Convolutional layers with ReLU activation for feature extraction.
• MaxPooling2D: Pooling layers to reduce spatial dimensions.
• BatchNormalization: Normalizes the output of the previous layer.
• Dropout: Regularization technique to prevent overfitting.
• Flatten: Converts 2D matrix data to a vector.
• Dense: Fully connected layers with ReLU and softmax activation for classification.

Model Compilation

• Compile Model: Use Adam optimizer and categorical cross-entropy loss function.

Data Augmentation

• ImageDataGenerator: Augments training data to improve model generalization.

Loading Data

• Load Data: Load training and validation data from directories.

Model Training

• Train Model: Train the model using the training set and validate using the validation set.

Saving the Model

• Save Model: Save the trained model architecture and weights.

2. Model Testing with GUI

To make the model user-friendly, I created a GUI using Tkinter for testing the model with new images.

Code Explanation

• Imports: Import necessary libraries for creating the GUI and loading the model.

GUI Setup

• GUI Elements: Create a window with buttons and labels for user interaction.
• Model Loading: Load the saved model and weights.
• Image Prediction: Load and preprocess the selected image, make predictions, and display the result.

GitHub Link: https://github.com/heerthiraja/Deep-Learning-Projects/tree/main/Leaf-Disease-prediction-DL-Project

Conclusion

This project demonstrates the application of computer vision and deep learning in agriculture. By detecting leaf diseases early, farmers can take timely actions to protect their crops. A robust CNN model and a user-friendly GUI make this solution practical and accessible.