KLASIFIKASI KUALITAS KAYU KELAPA MENGGUNAKAN ARSITEKTUR CNN
Abstract
The increase in the export volume of coconut logs, which are materials that can efficiently substitute for conventional wood, demands that the quality of coconut wood classified quickly. However, due to the limitations of a grader as a human being, it is necessary to have assistance from machines or technology that can classify coconut wood quickly. Techniques that used for rapid classification can use computer visualization. Convolutional Neural Network (CNN) with the right architecture makes this method able to recognize and detect objects well, which influenced by computerized factors, large datasets, and techniques to train deeper networks. This study uses five types of CNN architecture, AlexNet, GoogLeNet, ResNet101, ResNet18, and ResNet50. The research results obtained for the classification of the quality of coconut wood using images show that the GoogLeNet architecture has the best classification performance among other architectures. GoogLeNet gets result with an average accuracy of 84.89% in each layer, followed by RestNet101 architecture with an average accuracy of 78.41%, RestNet50 with an average accuracy of 77.18%, RestNet18 with an average accuracy of 72.94% and the lowest accuracy performance among other architectures obtained by AlexNet with an average accuracy of 65.84%.
Keywords: Classification, Coconut Wood, Computer Visualization Techniques, CNN
Meningkatnya volume ekspor kayu kelapa yang merupakan bahan pengganti kayu konvensional secara efisien menuntut klasifikasi kualitas kayu kelapa dengan cepat. Namun karena keterbatasan seorang grader sebagai manusia maka diperlukan bantuan mesin atau teknologi yang dapat mengklasifikasikan kayu kelapa dengan cepat. Teknik yang dapat digunakan untuk klasifikasi cepat dapat menggunakan teknik visualisasi komputer. Convolutional Neural Network (CNN) dengan arsitektur yang tepat menjadikan metode ini mampu mengenali dan mendeteksi objek dengan baik, yang sebagian besar dipengaruhi oleh faktor komputerisasi, dataset yang besar, dan teknik untuk melatih jaringan yang lebih dalam. Penelitian ini menggunakan lima jenis arsitektur CNN yaitu, AlexNet, GoogLeNet, ResNet101, ResNet18, dan ResNet50. Hasil penelitian yang diperoleh untuk klasifikasi kualitas kayu kelapa menggunakan citra menunjukkan bahwa arsitektur GoogLeNet memiliki performansi klasifikasi terbaik diantara arsitektur lainnya. GoogLeNet mendapatkan hasil dengan rata-rata akurasi 84,89% pada setiap lapisan, disusul arsitektur RestNet101 dengan akurasi rata-rata 78,41%, RestNet50 dengan akurasi rata-rata 77,18%, RestNet18 dengan akurasi rata-rata 72,94% dan kinerja akurasi terendah di antara arsitektur lainnya diperoleh AlexNet dengan akurasi rata-rata 65,84%.
Kata kunci: Klasifikasi, Kayu Kelapa, Teknik Visualisasi Komputer, CNN
Full Text:
PDFReferences
J. Balfas, “Perlakuan Resin Pada Kayu Kelapa (Cocos Nucifera),” J. Penelit. Has. Hutan, vol. 25, no. 2, pp. 108–118, 2007.
D. Purwanto, “Finishing Kayu Kelapa (Cocos Nucifera L.) Untuk Bahan Interior Ruangan,” J. Ris. Ind. Has. Hutan, vol. 3, no. 2, pp. 32–37, 2011.
I. Yuwono, R. Pramunendar, P. Andono, and R. A. Subandi, “The Quality Determination Of Coconut Wood Density Using Learning Vector Quantization,” vol. 57, pp. 82–87, Nov. 2013.
H. Hendriyana and Y. H. Maulana, “Identification of Types of Wood using Convolutional Neural Network with Mobilenet Architecture,” J. RESTI Rekayasa Sist. Dan Teknol. Inf., vol. 4, no. 1, pp. 70–76, 2020.
G. A. Ruz, P. A. Estevez, and C. A. Perez, “A Neurofuzzy Color Image Segmentation Method for Wood Surface Defect Detection,” For. Prod. J., vol. 55, no. 4, pp. 52–58, 2005.
S. Gasim and Agus Harjoko, “Metode Identifikasi Jenis Kayu Berdasarkan Model Blok Citra Mikroskopis Penampang Melintang,” PhD Thesis, Universitas Gadjah Mada, Yogyakarta, 2014.
Fera Flaurensia, Tedy Rismawan, and Rahmi Hidayati, “Pengenalan Motif Batik Indonesia Menggunakan Deteksi Tepi Canny dan Template Matching,” Coding J. Komput. Dan Apl., vol. 4, no. 2, 2016.
Ratri Dwi Dkk Atmaja, Achmad Rizal, and Koredianto Usman, “Deteksi Jenis Kayu Citra Furniture Ukiran Jepara Menggunakan JST Backpropagation,” Konf. Nas. Sist. Inf. STMIK–STIKOM Bdg., 2012.
Husnul Khatimi and Yuslena Sari, “Otomatisasi Tingkat Kualitas Kayu Kelapa Menggunakan Genetic Algorithm,” INFO-Tek., vol. 20, no. 2, pp. 255–264, Desember 2019.
Yann LeCun, Léon Bottou, Yoshua Bengio, and Patrick Haffner, “Gradient-Based Learning Applied to Document Recognition,” Proc. IEEE, vol. 86, no. 11, pp. 2278–2324, 1998.
Alex Krizhevsky, Ilya Sutskever, and Geoffrey E. Hinton, “Imagenet classification with deep convolutional neural networks,” Adv. Neural Inf. Process. Syst., vol. 25, pp. 1097–1105, 2012.
Christian Szegedy, Vincent Vanhoucke, Sergey Ioffe, Jon Shlens, and Zbigniew Wojna, “Rethinking The Inception Architecture for Computer Vision,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 2818–2826.
Karen Simonyan and Andrew Zisserman, “Very Deep Convolutional Networks for Large-Scale Image Recognition,” ArXiv Prepr. ArXiv14091556, 2014.
DOI: http://dx.doi.org/10.20527/klik.v8i1.370
Copyright (c) 2021 KLIK - KUMPULAN JURNAL ILMU KOMPUTER
This work is licensed under a Creative Commons Attribution 4.0 International License.
This work is licensed under a Creative Commons Attribution 4.0 International License. View My Stats