Klasifikasi Kematangan Stroberi Menggunakan OpenCV, CNN EfficientNet-B0, dan Grad-CAM Berbasis Web

Dion Arifin; Tomi Tri Sujaka; I Nyoman  Switrayana

doi:10.31004/riggs.v5i2.12873

Authors

Dion Arifin Universitas Bumigora
Tomi Tri Sujaka Universitas Bumigora
I Nyoman Switrayana Universitas Bumigora

DOI:

https://doi.org/10.31004/riggs.v5i2.12873

Keywords:

Stroberi, Pengolahan Citra Digital, Convolutional Neural Network, EfficientNet-B0, Grad-CAM

Abstract

Penilaian tingkat kematangan buah stroberi secara manual bersifat subjektif, memerlukan pengalaman, dan memakan waktu, sehingga dibutuhkan sistem klasifikasi otomatis yang akurat, transparan, serta mudah diakses. Penelitian ini mengembangkan sistem klasifikasi tingkat kematangan stroberi berbasis web dengan memanfaatkan Open Source Computer Vision Library (OpenCV) untuk pemrosesan citra dan Convolutional Neural Network berarsitektur EfficientNet-B0 melalui transfer learning dengan strategi partial fine-tuning. Dataset terdiri atas 677 citra stroberi yang diperoleh dari Kecamatan Sembalun, Kabupaten Lombok Timur, dan repositori publik, kemudian dikelompokkan ke dalam kelas “Dapat Dipetik” dan “Tidak Dapat Dipetik”. Data dibagi menggunakan stratified random sampling melalui tiga skenario rasio latih, validasi, dan uji, yaitu 80:10:10, 70:15:15, dan 60:20:20. Transparansi keputusan model diwujudkan melalui Gradient-weighted Class Activation Mapping (Grad-CAM) yang diproses bersama OpenCV untuk menghasilkan heatmap dan overlay. Sistem dikembangkan menggunakan FastAPI, dikemas melalui Docker, dan disebarkan pada Hugging Face Spaces. Hasil pengujian menunjukkan bahwa rasio 80:10:10 menghasilkan kinerja paling stabil dengan akurasi validasi akhir sebesar 97,79% dan F1-score rata-rata tertimbang sebesar 97,79%. Visualisasi Grad-CAM menunjukkan bahwa model berfokus pada tekstur dan pigmentasi buah, bukan pada latar belakang. Temuan ini menunjukkan bahwa integrasi EfficientNet-B0, Grad-CAM, OpenCV, dan layanan web mampu menyediakan klasifikasi kematangan stroberi yang akurat, dapat dijelaskan, responsif, dan berpotensi mendukung pengambilan keputusan panen.

Downloads

Download data is not yet available.

References

[1] S. Afrin, M. Gasparrini, T. Y. Forbes-Hernandez, et al., “Promising Health Benefits of the Strawberry: A Focus on Clinical Studies,” J. Agric. Food Chem., vol. 64, no. 22, pp. 4435–4449, 2016.

[2] I. Setiawan, A. C. Tridakusumah, M. A. Heryanto, et al., “Socio-Economic Factors Influencing Strawberry Farming Development in West Java, Indonesia,” BIO Web Conf., vol. 225, p. 07005, 2026.

[3] E. S. Hapsari, Tajidan, and Muktasam, “Social, Economic, and Environmental Factors in Determining the Sustainability of Tourist Attractions: The Case of Strawberry Agro-Tourism in Sembalun District, Indonesia,” Asian J. Agric. Hortic. Res., vol. 12, no. 3, pp. 13–26, 2025.

[4] Z. Mi and W. Q. Yan, “Strawberry Ripeness Detection Using Deep Learning Models,” Big Data Cogn. Comput., vol. 8, no. 8, p. 92, 2024.

[5] W. Hamri and W. M. S. Jabbar, “Classification of Strawberry Ripeness Stages Using Deep Learning,” Evol. Electr. Electron. Eng.

[6] B. Yang and Y. Xu, “Applications of Deep-Learning Approaches in Horticultural Research: A Review,” Hortic. Res., vol. 8, no. 1, p. 123, 2021.

[7] J. Park, J. Shin, R. Kim, et al., “Accelerating Strawberry Ripeness Classification Using a Convolution-Based Feature Extractor along with an Edge AI Processor,” Electronics, vol. 13, no. 2, p. 344, 2024.

[8] J. J. K. Chai, J.-L. Xu, and C. O’Sullivan, “Real-Time Detection of Strawberry Ripeness Using Augmented Reality and Deep Learning,” Sensors, vol. 23, no. 17, p. 7639, 2023.

[9] W. J. von Eschenbach, “Transparency and the Black Box Problem: Why We Do Not Trust AI,” Philos. Technol., vol. 34, no. 4, pp. 1607–1622, 2021.

[10] M. Tan and Q. V. Le, “EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks,” arXiv:1905.11946, 2019. doi: 10.48550/arXiv.1905.11946.

[11] É. K. Hamaguti and F. A. Breve, “Classification of Strawberry Maturity Level from Images Using Supervised Classifiers and Convolutional Neural Networks as Feature Extractors,” in Lecture Notes in Computer Science, Springer Nature Switzerland, 2025.

[12] R. R. Selvaraju, M. Cogswell, A. Das, R. Vedantam, D. Parikh, and D. Batra, “Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization,” in Proc. IEEE ICCV, 2017, pp. 618–626.

[13] G. Bradski, “The OpenCV Library,” Dr. Dobb’s J. Softw. Tools, 2000.

[14] M. Sandler, A. Howard, M. Zhu, A. Zhmoginov, and L.-C. Chen, “MobileNetV2: Inverted Residuals and Linear Bottlenecks,” in Proc. IEEE CVPR, 2018, pp. 4510–4520.

[15] J. Hu, L. Shen, and G. Sun, “Squeeze-and-Excitation Networks,” in Proc. IEEE CVPR, 2018, pp. 7132–7141.

[16] P. Ramachandran, B. Zoph, and Q. V. Le, “Searching for Activation Functions,” arXiv:1710.05941, 2018.

[17] C. Shorten and T. M. Khoshgoftaar, “A Survey on Image Data Augmentation for Deep Learning,” J. Big Data, vol. 6, no. 1, pp. 1–48, 2019.