Jurnal Elkomika (Jan 2024)

Sistem Pencegahan Illegal Fishing di Laut Batam menggunakan YOLOv7 berbasis Notifikasi Telegram

  • MUHAMMAD ABRAR MASRIL,
  • DEOSA PUTRA CANIAGO

DOI
https://doi.org/10.26760/elkomika.v12i1.175
Journal volume & issue
Vol. 12, no. 1

Abstract

Read online

ABSTRAK Pulau Batam menjadi salah satu pulau indonesia terluar yang berbatasan langsung dengan negara tetangga. Penerapan YOLOv7 untuk mendeteksi kapal di laut Batam mampu mendeteksi objek kapal dengan hasil pengujian setelah melakukan training 100 epoch menghasilkan nilai precision sebesar 1.00 dan nilai confidence 0.882 menunjukkan tingkat kepercayaan hasil deteksi yang tinggi pada model YOLOv7. Hasil skor F1 sebesar 0.99 pada confidence 0.729 menunjukkan hasil bahwa model ini menghasilkan tingkat akurasi yang tinggi dalam menemukan objek. Berdasarkan hasil evaluasi menggunakan confusion matrix menunjukkan hasil akurasi yang tinggi dari setiap class pada model YOLOv7 yaitu Ferry 93%, KapalNelayanIndo 85%, KapalNelayanMalaysia 89%, KapalNelayanThailand 91%, KapalNelayanVietnam 82%, Speedboat 94%, dan Tanker 83%. Hasil pengujian aplikasi website yang terintegrasi dengan YOLOv7 dan bot Telegram menghasilkan website yang mampu mendeteksi objek dan mengirimkan notifikasi sehingga diharapkan mampu mencegah illegal fishing. Kata kunci: Deteksi, Deep Learning, Kapal, Telegram, YOLOv7 Â ABSTRACT Batam Island is one of Indonesia's outermost islands bordering neighboring countries. The application of YOLOv7 to detect ships in the Batam Sea was able to detect ship objects with test results after carrying out 100 epoch training resulting in a precision value of 1.00 and a confidence value of 0.882 indicating a high level of confidence in the detection results in the YOLOv7 model. The F1 score of 0.99 at confidence 0.729 shows that this model produces high accuracy in finding objects. Based on the evaluation results using the confusion matrix, it shows high accuracy results for each class in the YOLOv7 model, namely Ferry 93%, Indonesian Fisherman's Ship 85%, Malaysian Fisherman's Ship 89%, Thai Fisherman's Ship 91%, Vietnamese Fisherman's Ship 82%, Speedboat 94%, and Tanker 83%. The test results of the website application integrated with YOLOv7 and Telegram bot resulted in a website that is able to detect objects and send notifications so that it is expected to be able to prevent illegal fishing. Keywords: Detection, Deep Learning, Ship, Telegram, YOLOv7

Keywords