KLASIFIKASI CITRA ASET BERBASIS CNN DAN METADATA EXIF UNTUK REKAP DATA LAPANGAN DI PT.XYZ
DOI:
https://doi.org/10.54314/7a6qj068Keywords:
CNN, image classification, EXIF metadata, field data recapitulation, assetAbstract
Abstract: This study aims to design and build an asset and non-asset photo classification system based on a Convolutional Neural Network (CNN), integrated with EXIF metadata extraction, to automate field data recapitulation at PT.XYZ using the Agile method. The underlying problem is that field photos are recorded one by one, which is time-consuming and error-prone, and not every photo in a survey folder is an asset photo. The dataset consists of 994 images divided into training, validation, and test data. Images are processed through a 3×3 median filter, resizing to 224×224 pixels, and normalization, then classified by a CNN model with three convolution blocks. Photos classified as assets have their timestamp and coordinates extracted into an Excel recap and a KML file. Testing on 150 test images yielded an accuracy of 96.67%, precision of 98.61%, recall of 94.67%, and F1-score of 96.60%, while on 75 field images the accuracy reached 78.67% with precision of 100% and recall of 77.14%. The CNN model can screen asset photos and automate field data recapitulation, although recall on field data still needs improvement.
Keywords: CNN; image classification; EXIF metadata; field data recapitulation; asset
Abstrak: Penelitian ini bertujuan merancang dan membangun sistem klasifikasi foto aset dan non-aset berbasis Convolutional Neural Network (CNN) yang terintegrasi dengan ekstraksi metadata EXIF untuk mengotomatisasi rekap data lapangan pada PT.XYZ dengan menggunakan metode Agile. Permasalahan yang melatarbelakangi adalah pencatatan foto lapangan yang dilakukan satu per satu sehingga memakan waktu dan rentan kesalahan, serta tidak semua foto dalam folder survei merupakan foto aset. Dataset terdiri atas 994 citra yang dibagi menjadi data latih, validasi, dan uji. Citra diproses melalui median filter 3×3, resize 224×224 piksel, dan normalisasi, kemudian diklasifikasikan oleh model CNN tiga blok konvolusi. Foto yang terklasifikasi sebagai aset diekstraksi waktu dan koordinatnya menjadi rekap Excel dan berkas KML. Hasil pengujian pada 150 citra uji menunjukkan akurasi 96,67%, precision 98,61%, recall 94,67%, dan F1-score 96,60%, sedangkan pada 75 citra lapangan akurasi mencapai 78,67% dengan precision 100% dan recall 77,14%. Disimpulkan bahwa model CNN mampu menyaring foto aset dan mengotomatisasi rekap data lapangan, namun recall pada data lapangan masih perlu ditingkatkan.
Kata kunci: CNN; klasifikasi citra; metadata EXIF; rekap data lapangan; aset
Downloads
References
Apriliani, A., Hijjayanti, K., & Suhairoh, S. (2020). Analisis keaslian citra dengan menggunakan EXIF metadata. CESS (Journal of Computer Engineering, System and Science), 5(1), 84–88. https://doi.org/10.24114/cess.v5i1.15600
Camera & Imaging Products Association, & Japan Electronics and Information Technology Industries Association. (2016). Exchangeable image file format for digital still cameras: Exif Version 2.31 (CIPA DC-008-2016; JEITA CP-3451D).
Harahap, L. S., & Kembaren, M. F. H. (2025). Klasifikasi objek menggunakan Convolutional Neural Network pada citra satelit. RIGGS: Journal of Artificial Intelligence and Digital Business, 4(2), 4663–4667. https://doi.org/10.31004/riggs.v4i2.1306
Ioffe, S., & Szegedy, C. (2015). Batch normalization: Accelerating deep network training by reducing internal covariate shift. Proceedings of the 32nd International Conference on Machine Learning, PMLR 37, 448–456.
Kingma, D. P., & Ba, J. (2015). Adam: A method for stochastic optimization. Proceedings of the 3rd International Conference on Learning Representations. https://arxiv.org/abs/1412.6980
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). ImageNet classification with deep convolutional neural networks. Advances in Neural Information Processing Systems, 25, 1097–1105.
LeCun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998). Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11), 2278–2324. https://doi.org/10.1109/5.726791
Prastyo, A. E. (2025). Penerapan MobileNetV1 untuk identifikasi otomatis varietas beras melalui gambar smartphone [Skripsi, Universitas Mercu Buana]. https://repository.mercubuana.ac.id/96682
Powers, D. M. W. (2011). Evaluation: From precision, recall and F-measure to ROC, informedness, markedness and correlation. Journal of Machine Learning Technologies, 2(1), 37–63.
Srivastava, N., Hinton, G., Krizhevsky, A., Sutskever, I., & Salakhutdinov, R. (2014). Dropout: A simple way to prevent neural networks from overfitting. Journal of Machine Learning Research, 15(56), 1929–1958.
Wijaya, A. C., Mahbubi, A. B., Januarta, M. F., Syabila, T., & Zakaria, D. (2025). Analisis sistematis algoritma Convolutional Neural Network untuk klasifikasi gambar bokeh dan blur: Tinjauan literatur. Jurnal JTIK, 9(4). https://doi.org/10.35870/jtik.v9i4.3953
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Wilhelmus Gulo, Masdiana Sagala (Author)
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
