Wahab Sya'roni, Yahya Auliya' Abdillah, Samsul Arifin, Rayhan Hibatullah

Klasifikasi Penyakit Daun Sawi Menggunakan VGG19 Berbasis Citra Digital

Introduction

Klasifikasi penyakit daun sawi menggunakan vgg19 berbasis citra digital. Klasifikasi penyakit daun sawi otomatis menggunakan VGG19 berbasis citra digital. Raih akurasi 96% untuk deteksi penyakit tanaman. Potensial untuk sistem IoT pertanian guna bantu petani.

Abstract

Agricultural productivity is greatly influenced by plant health, including mustard greens (Brassica rapa), which are prone to leaf diseases and have high economic value. This study aims to develop a digital image-based classification system for mustard leaf diseases using a deep learning approach, particularly the Convolutional Neural Network (CNN) VGG19 architecture, and to compare its performance with ResNet50 and VGG16 models. The dataset used consists of 999 images divided into two classes: healthy mustard leaves and diseased mustard leaves. The images were processed through preprocessing steps (resized to 224×224 and normalized), then split into training, validation, and testing sets (80:10:10). The VGG19 architecture was customized with additional layers such as Global Average Pooling and Dense layers, and trained for 50 epochs with a configuration of 32 filters, a dropout rate of 0.5, and a learning rate of 0.0003. The results showed that the VGG19 model achieved the highest validation accuracy of 96%, followed by VGG16 with 95%, and ResNet50 with 74%. Evaluation using a confusion matrix demonstrated that VGG19 exhibited the most stable and accurate classification performance in recognizing both classes. These findings reinforce the potential of VGG19 for developing automated and real-time plant disease detection systems. Furthermore, this study opens up opportunities for integration into agricultural Internet of Things (IoT) systems for continuous plant health monitoring, thereby assisting farmers in making faster and more accurate preventive decisions.

Review

This paper presents a timely and relevant study on the classification of mustard leaf diseases using digital images, employing deep learning techniques. The authors leverage the VGG19 Convolutional Neural Network architecture and perform a comparative analysis against VGG16 and ResNet50. The research addresses a significant agricultural challenge by aiming to improve plant health monitoring and productivity, particularly for mustard greens, a crop of high economic value. The methodological approach, which includes preprocessing, dataset splitting, and customized model training, appears sound, and the reported high accuracy of VGG19 for binary classification demonstrates promising potential for automated disease detection. The study's strength lies in its systematic comparison of three established CNN architectures, clearly demonstrating VGG19's superior performance, achieving a 96% validation accuracy. While the use of a dataset of 999 images is a good starting point, the limitation to only two classes (healthy and diseased) simplifies the problem, and a more diverse dataset encompassing multiple disease types would enhance the model's robustness and real-world applicability. It would also be insightful to further discuss the significant performance gap observed with ResNet50 (74%), as this architecture is generally known for its high performance; potential reasons related to hyperparameter tuning, specific dataset characteristics, or model suitability for the binary classification task could be explored. The detailed description of VGG19 customization, including Global Average Pooling and Dense layers, adds credibility to the model development. Overall, this research provides a valuable contribution to the field of precision agriculture, laying a solid foundation for automated mustard leaf disease detection. The identified potential for integration into IoT systems for continuous monitoring highlights a clear path towards practical application, enabling farmers to make faster and more accurate preventive decisions. For future work, expanding the dataset to include a wider variety of diseases, varying severity levels, and images captured under diverse environmental conditions would be crucial. Investigating the computational efficiency of the VGG19 model for resource-constrained IoT devices, and perhaps exploring lighter-weight architectures or model compression techniques, could further enhance its practical utility. This study effectively demonstrates the capabilities of deep learning in addressing agricultural challenges and offers exciting avenues for subsequent research and development.

Full Text

Comments

You need to be logged in to post a comment.