IMPLEMENTASI DETEKSI KENDARAAN MENGGUNAKAN METODE YOLOV5

JAYANTI, DWI (2023) IMPLEMENTASI DETEKSI KENDARAAN MENGGUNAKAN METODE YOLOV5. S1 thesis, Universitas Mercu Buana Jakarta.

Preview	Text (COVER) 01 COVER.pdf Download (310kB) | Preview
Preview	Text (ABSTRAK) 02 ABSTRAK.pdf Download (126kB) | Preview
	Text (BAB I) 03 BAB 1.pdf Restricted to Registered users only Download (67kB)
	Text (BAB II) 04 BAB 2.pdf Restricted to Registered users only Download (251kB)
	Text (BAB III) 05 BAB 3.pdf Restricted to Registered users only Download (461kB)
	Text (BAB IV) 06 BAB 4.pdf Restricted to Registered users only Download (1MB)
	Text (BAB V) 07 BAB 5.pdf Restricted to Registered users only Download (57kB)
	Text (DAFTAR PUSTAKA) 08 DAFTAR PUSTAKA.pdf Restricted to Registered users only Download (38kB)

Abstract

Nowadays, information technology is progressing and developing rapidly, which has a huge impact on various aspects of life. One area that has seen an increase in research is artificial intelligence, particularly in object detection in images. This research focuses on the implementation of an object detection system for vehicles such as cars, buses, and trucks, using the You Only Look Once (YOLO) method. The results show that the utilization of the You Only Look Once (YOLO) method in vehicle detection provides good results. In tests using 317 images in the dataset, the model obtained a [email protected] value of 99.5%, with a batch size of 15 and underwent 200 epochs on the recall value. The average recall value peaked at 1 at a confidence level of 0.00. In addition, during the daytime, the model achieved an accuracy rate of 74.3%. The findings of this study show that the YOLO method is reliable for vehicle detection with satisfactory results. The development of this kind of object detection technology has great potential in a variety of applications, including traffic security, city surveillance, and other fields that require visual analysis and object recognition quickly and accurately. Keywords: Object Detection, Google Colab, Python, You Only Look Once (YOLO), GitHub. Saat ini, teknologi informasi mengalami kemajuan dan perkembangan pesat, yang memiliki dampak besar pada berbagai aspek kehidupan. Salah satu bidang yang telah mengalami peningkatan dalam penelitian adalah kecerdasan buatan, khususnya dalam deteksi objek pada gambar. Penelitian ini fokus pada implementasi sistem deteksi objek untuk kendaraan seperti mobil, bus, dan truk, menggunakan metode You Only Look Once (YOLO). Hasil penelitian menunjukkan bahwa pemanfaatan metode You Only Look Once (YOLO) dalam deteksi kendaraan memberikan hasil yang baik. Dalam pengujian menggunakan 317 citra pada dataset, model memperoleh nilai mAP@0,5 sebesar 99,5%, dengan batch size 15 dan menjalani 200 epochs terhadap nilai recall. Nilai rata-rata recall mencapai puncaknya pada nilai 1 pada tingkat kepercayaan (confidence) 0,00. Selain itu, pada siang hari, model mencapai tingkat akurasi sebesar 74,3%. Temuan dari penelitian ini menunjukkan bahwa metode YOLO dapat diandalkan untuk deteksi kendaraan dengan hasil yang memuaskan. Perkembangan teknologi deteksi objek semacam ini memiliki potensi besar dalam berbagai aplikasi, termasuk keamanan lalu lintas, pengawasan kota, dan bidang lain yang memerlukan analisis visual dan pengenalan objek secara cepat dan akurat. Kata Kunci : Deteksi Objek, Google Colab, Python, You Only Look Once (YOLO), GitHub.

Actions (login required)

View Item