Priyandra, Rendi (2025) Implementasi Convolutional Neural Network Dengan Tensorflow Untuk Deteksi Penyakit Pada Daun Pisang. Other thesis, Universitas Islam Riau.

Text 203510427.pdf - Submitted Version Restricted to Registered users only Download (7MB) | Request a copy

Abstract

Detection of diseases on banana leaves is an important challenge in agriculture that often results in reducing productivity and economic losses for farmers. Traditional methods of identifying these diseases generally require direct observation by humans, which is often time-consuming, laborious, and prone to errors. This research aims to develop an automated system using Convolutional Neural Network (CNN) model to predict diseases on banana leaves. Three CNN architectures, namely MobileNetV2, DenseNet121, and NasNet Mobile, were implemented using TensorFlow in this study. The dataset used consists of 5400 samples, which were obtained from Kaggle (BananaLSD) and banana plantations in Kuantan Singingi Regency. The dataset includes images of banana leaves infected with Sigatoka and Cordana diseases, as well as images of healthy leaves. Data augmentation techniques were applied to increase the variety of the dataset. The results showed that the DenseNet121 model achieved the highest accuracy of 99.07%, with an error rate of 0.93%, precision of 99.08%, recall of 99.08%, and F1-score of 99.08%, making it the most reliable model compared to the other two models. This research contributes a practical approach to disease prediction that can be integrated into a mobile application, enabling farmers to identify banana leaf diseases quickly and accurately, as well as providing information on followup actions after disease diagnosis, such as treatment and disease control.

Actions (login required)

View Item