Implementasi MobileNetV2 dengan Teknik Augmentasi Data untuk Klasifikasi Penyakit Daun Cabai Berbasis Website
DOI:
https://doi.org/10.36312/txr3y475Keywords:
Augmentasi Data; Klasifikasi Citra; Mobilenetv2; Penyakit Daun Cabai; Validasi Model; WebsiteAbstract
Penelitian ini bertujuan mengembangkan model klasifikasi penyakit daun cabai berbasis MobileNetV2 yang diintegrasikan ke dalam prototipe website sebagai alat bantu diagnosis awal. Permasalahan utama yang diangkat adalah kemiripan gejala visual antarpenyakit daun cabai, keterbatasan identifikasi manual, dan kebutuhan model yang ringan untuk implementasi praktis. Dataset terdiri atas 500 citra awal dari lima kelas, yaitu daun sehat, daun keriting, daun kuning, bercak daun, dan embun tepung. Data diperluas melalui augmentasi rotasi, shear, zoom, dan horizontal flip sehingga diperoleh 2.200 citra, kemudian dibagi menjadi data latih 80%, validasi 10%, dan uji 10%. Model dilatih menggunakan arsitektur MobileNetV2 dengan optimizer Adam, batch size 32, dan 20 epoch. Hasil menunjukkan bahwa augmentasi data meningkatkan akurasi model dari 74,00% menjadi 99,09%. Pada data uji, model mencapai akurasi 98,64% dengan loss 0,0363. Classification report memperlihatkan performa tinggi pada seluruh kelas, meskipun masih terdapat kesalahan minor pada kelas daun keriting dan embun tepung. Sistem berbasis website juga berhasil menjalankan fungsi registrasi, login, unggah citra, klasifikasi, konsultasi AI, dan profil pengguna melalui pengujian black-box. Secara kritis, hasil ini menunjukkan potensi MobileNetV2 sebagai model efisien untuk klasifikasi penyakit daun cabai, namun validasi eksternal pada kondisi lapangan yang lebih beragam tetap diperlukan sebelum sistem digunakan sebagai alat diagnosis pertanian berskala luas.
This study aims to develop a MobileNetV2-based chili leaf disease classification model integrated into a website prototype as an early diagnostic support tool. The central problem concerns the visual similarity among chili leaf disease symptoms, the limitations of manual identification, and the need for a lightweight model suitable for practical deployment. The dataset consisted of 500 original images from five classes, namely healthy leaf, curly leaf, yellow leaf, leaf spot, and powdery mildew. Data augmentation using rotation, shear, zoom, and horizontal flipping expanded the dataset to 2,200 images, which were divided into 80% training, 10% validation, and 10% testing subsets. The model was trained using the MobileNetV2 architecture with the Adam optimizer, a batch size of 32, and 20 epochs. The results show that data augmentation improved model accuracy from 74.00% to 99.09%. On the test set, the model achieved 98.64% accuracy with a loss value of 0.0363. The classification report indicates strong performance across all classes, although minor misclassifications were found in curly leaf and powdery mildew classes. The web-based system also successfully supported registration, login, image upload, classification, AI consultation, and user profile functions through black-box testing. Critically, these findings demonstrate the potential of MobileNetV2 as an efficient model for chili leaf disease classification, but external validation under more diverse field conditions is still necessary before the system can be used as a large-scale agricultural diagnostic tool.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Muhammad Irzan Ali

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Authors who publish with Journal of Authentic Research agree to the following terms:
- For all articles published in Journal of Authentic Research, copyright is retained by the authors. Authors give permission to the publisher to announce the work with conditions. When the manuscript is accepted for publication, the authors agrees to implement a non-exclusive transfer of publishing rights to the journals.
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution-ShareAlike 4.0 International License that allows others to share the work with an acknowledgment of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgment of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.