Deteksi Objek Kereta Api menggunakan Metode Faster R-CNN dengan Arsitektur VGG 16

JASMAN PARDEDE; HENDRI HARDIANSAH

doi:10.26760/mindjournal.v7i1.21-36

Deteksi Objek Kereta Api menggunakan Metode Faster R-CNN dengan Arsitektur VGG 16

JASMAN PARDEDE, HENDRI HARDIANSAH

Sari

ABSTRAK

Kereta merupakan sebuah alat transportasi umum yang sering digunakan oleh masyarakat untuk berpergian dari kota asal ke kota tujuan. Mereka membutuhkan akan sarana transportasi umum untuk mempermudah aktifitas mereka. Namun kecelakaan di persimpangan jalan raya yang terlintasi oleh kereta api memiliki angka yang cukup besar akibat kelalaian dari petugas untuk menutup palang pintu kereta api. Maka dari itu penelitian ini dibuat agar mengetahui keberadaan kereta api berdasarkan jarak dan tingkat cahayanya dari siang sampai malam hari. Sistem dibangun menggunakan metode Faster RCNN dengan model arsitektur VGG16 untuk mengetahui keberadaan objek kereta api antara lokomotif dan gerbong berdasarkan tingkat cahaya dan jarak terhadap objek. Setelah dilakukan pengujian dengan jarak paling dekat Â±2 meter sampai Â±250 meter, diperoleh rata-rata akurasi untuk lokomotif sebesar 79,09%, dan akurasi untuk gerbong sebesar 97,05%. Sistem memperoleh keakurasian deteksi terhadap objek rata-rata akurasi deteksi objek lokomotif sebesar 86,40%, dan rata-rata akurasi deteksi objek gerbong sebesar 97,23%.

Kata kunci: Deteksi Objek, Faster RCNN, VGG, Kereta Api, Jarak, Lux

ABSTRACT

Railway is a public transportation that is often used by the public to travel from the home town to the destination city. They need public transportation to facilitate their activities. But accidents at the intersection of the highway crossed by the train has a considerable number due to the negligence of the officer to close the railway stopbars. Therefore, this study was made to know the existence of trains based on their distance and light level from day to night. The system was built using the Faster RCNN method with the VGG16 architectural model to determine the existence of railway objects between locomotives and carriages based on the level of light and distance to the object. After testing with the closest distance of Â±2 meters to Â±250 meters, obtained an average accuracy for locomotives of 79.09%, and accuracy for carriages of 97.05%. The system obtained accuracy of detection of objects with an average detection accuracy of locomotive objects of 86.40%, and an average detection accuracy of car objects of 97.23%.

Keywords: Object Detection, Faster RCNN, VGG, Railway, Distance, Lux

Teks Lengkap:

PDF

Referensi

Benuwa, B., Zhan, Y., Ghansah, B. (2016). A Review of Deep Machine Learning. International Journal of Engineering Research in Africa. 10.4028/www.scientific.net/JERA.24.124

Girshick, R. (2015). ICCV paper the Open Access Version by Computer Vision Foundation. Fast R-CNN.

He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual learning for image recognition. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2016-December, 770â€“778. https://doi.org/10.1109/CVPR.2016.90

Huang, H., Zhou, H., Yang, X., Zhang, L., Qi, L., & Zang, A. Y. (2019). Faster R-CNN for marine organisms detection and recognition using data augmentation. Neurocomputing, 337(xxxx), 372â€“384. https://doi.org/10.1016/j.neucom.2019.01.084

KBBI. 2020. â€œKereta Apiâ€ : https://kbbi.kemdikbud.go.id/entri/Kereta%20Api.

Khan, A., Sohail, A., Zahoora, U., & Qureshi, A. S. (2020). A survey of the recent architectures of deep convolutional neural networks. Artificial Intelligence Review, 53(8), 5455â€“5516. https://doi.org/10.1007/s10462-020-09825-6

Khurana, K., & Awasthi, R. (2013). Techniques for Object Recognition in Images and Multi-Object Detection. International Journal of Advanced Research in Computer Engineering & Technology, 2(4), 1383â€“1388.

Lee, C., Kim, H. J., & Oh, K. W. (2016). Comparison of faster R-CNN models for object detection. International Conference on Control, Automation and Systems, 0(Iccas), (pp. 107â€“110). https://doi.org/10.1109/ICCAS.2016.7832305

Mathworks. 2020. â€œWhat Is Object Detection?â€ https://www.mathworks.com/discovery/object-detection.htm.

Mathworks. 2018. â€œWhat Is Deep Learning?â€ https://www.mathworks.com/discovery/deep-learning.html.

Nagataries, D., Hardiristanto, S., Purnomo, M. H., & Klasik, A. A. G. (2012). Deteksi Objek pada Citra Digital Menggunakan Algoritma Genetika untuk Studi Kasus Sel Sabit. Journal of Electrical Engineering.

Ren, S., He, K., Girshick, R., & Sun, J. (2017). Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks. IEEE Transactions on Pattern Analysis and Machine Intelligence, 39(6), 1137â€“1149. https://doi.org/10.1109/TPAMI.2016.2577031

Roh, M. C., & Lee, J. Y. (2017). Refining faster-RCNN for accurate object detection. Proceedings of the 15th IAPR International Conference on Machine Vision Applications, MVA 2017, (pp. 514â€“517). https://doi.org/10.23919/MVA.2017.7986913

Rosebrock, A. (2016). Intersection over Union (IoU) for object detection. Diambil kembali dari PYImageSearch: https://www.pyimagesearch.com/2016/11/07/intersectionover-union-iou-for-object-detect

Russakovsky, O., Deng, J., Su, H., Krause, J., Satheesh, S., Ma, S., â€¦ Fei-Fei, L. (2015). ImageNet Large Scale Visual Recognition Challenge. International Journal of Computer Vision, 115(3), 211â€“252. https://doi.org/10.1007/s11263-015-0816-y

Srijakkot, K., Kanjanasurat, I., Wiriyakrieng, N., & Benjangkaprasert, C. (2020). The comparison of Faster R-CNN and Atrous Faster R-CNN in different distance and light condition. Journal of Physics: Conference Series, 1457(1), 0â€“6. https://doi.org/10.1088/1742-6596/1457/1/012015

Tempo, 2019. â€œKAI Catat 655 Kecelakaan di Perlintasan Kereta 2 Tahun Terakhirâ€ https://bisnis.tempo.co/read/1244503/kai-catat-655-kecelakaan-di-perlintasan-kereta-2-tahun-terakhir.

Y. Lecun, Y. Bengio, dan G. Hinton. (2015). Deep learning. Nature, 521(7553), 436â€“444.

DOI: https://doi.org/10.26760/mindjournal.v7i1.21-36