Plant Disease Detection with Machine Learning
Mansi Dupte, Khushi Moolya, Sakshi Maurya, Shweta Phanse, Prof. Sarita Bopalkar
Students, Department of Computer Science and Engineering (IOT and Cybersecurity Including Blockchain Technology), Smt. Indira Gandhi College of Engineering, Navi Mumbai, India
Guide, Department of Computer Science and Engineering (IOT and Cybersecurity Including Blockchain Technology), Smt. Indira Gandhi College of Engineering, Navi Mumbai, India
Abstract: Plant diseases significantly impact agricultural productivity, necessitating early and accurate detection methods. Traditional manual inspection techniques are time-consuming and prone to inaccuracies. In this study, we propose a convolutional neural network (CNN)-based model for plant disease detection, with a specific fopcus on tomato leaf diseases. The dataset used comprises 11,000 images sourced from the PlantVillage database, encompassing multiple disease classes and healthy samples. The proposed CNN model consists of convo2D, batch normalization, max pooling, global average pooling, dense and dropout convolutional layers, and 3 fully connected layers, and employs a learning rate of 0.0002, trained over 50 epochs. Compared with traditional machine learning approaches such support vector machines (SVMs) and random forests (RFs), our model has a superior training accuracy of 98.15% and a validation accuracy of 96.2%, significantly reducing the number of false positives. The model incorporates advanced techniques such as batch normalization and dropout regularization, ensuring robustness and generalizability. The experimental results demonstrate the model's efficacy in accurately diagnosing tomato diseases, providing a reliable tool for precision agriculture. The proposed system not only automates the disease identification process but also offers treatment recommendations, thus contributing to enhanced crop health management.
Keywords: Plant Disease Detection, Convolutional Neural Network (CNN), Tomato Leaf Diseases, PlantVillage Dataset, Batch Normalization, Dropout Regularization, Precision Agriculture, Machine Learning (ML), Learning Rate, Accuracy Rate
