Prakoso, Yoga Adi (2026) Klasifikasi Penyakit Tanaman Gandum Menggunakan Metode Resnet50-V2. Other thesis, Universitas Islam Riau.

Text skripsi_213510356_watermark.pdf - Published Version Restricted to Registered users only Download (5MB) | Request a copy

Abstract

Wheat (Triticum aestivum L.) is an important food commodity, but its productivity often declines due to disease attacks such as Brown Rust, Loose Smut, Powdery Mildew, Septoria Leaf Blotch, and Yellow Rust. The disease identification process, which is still carried out manually, is time-consuming and depends on expert knowledge, so a fast and accurate automated method is needed. This study aims to implement an image-based wheat disease classification system using the ResNet50-V2 Convolutional Neural Network (CNN) architecture. The dataset consists of 841 wheat leaf images with six classes, which were augmented through static augmentation. The model was developed using a transfer learning approach with partial fine-tuning. Testing results show that the best configuration uses the AdamW optimizer with a learning rate of 1e-5 and 30 epochs, resulting in a test accuracy of 98.42%. The evaluation was conducted using a confusion matrix and accuracy, precision, recall, and F1-score metrics. The best model was implemented into a Flask-based website application to facilitate the wheat plant disease classification process.

Actions (login required)

View Item