Perancangan Automated Obstacle Detection System Pada Kapal Ikan 30 Gt Sebagai Early-Warning Tabrakan Pada Kapal

Putri, Novalyina Ghassani (2025) Perancangan Automated Obstacle Detection System Pada Kapal Ikan 30 Gt Sebagai Early-Warning Tabrakan Pada Kapal. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5018211013-Undergraduate_Thesis.pdf Restricted to Repository staff only Download (11MB) | Request a copy

Abstract

Pada akhir 2021 Komite Nasional Keselamatan Transportasi (KNKT) mencatat sekitar 31 persen dari kecelakaan kapal kapal sepanjang 2018-2020 terjadi pada kapal penangkap ikan. KNKT menyebutkan sekitar 100 orang per tahun yang meninggal dalam kecelakaan kapal ikan tersebut. Berdasarkan temuan KKP, sebagian besar kapal ikan tradisional tidak dilengkapi dengan alat navigasi, sehingga sering mengalami tabrakan dengan kapal lain, baik kapal besar maupun kapal ikan lainnya. Maka dari itu Tugas Akhir ini memiliki tujuan untuk membuat sebuah sistem deteksi kapal otomatis sehingga dapat meningkatkan keamanan dalam kegiatan operasional kapal ikan. Untuk mengembangkan sistem deteksi kapal otomatis perlu dilakukan studi terkait penghindaran tabrakan, sistem deteksi objek, machine learning. Dalam Tugas Akhir ini, algoritma YOLOv8 dipakai untuk melatih database gambar obstacle yang didapat dari bantuan Google Image Scraper. Terdapat empat klasifikasi obstacle yang terdeteksi yakni kapal ikan, tugboat, kapal keruk pasir, dan kapal cargo. Selanjutnya dilakukan perancangan hardware dengan menggunakan peralatan seperti Raspberry Pi sebagai mini pc sebagai otak dari pemograman automated obstacle detection system dan early-warning deteksi tabrakan pada kapal. Model terbaik dihasilkan setelah melakukan 120 kali iterasi dengan nilai matrix evaluasi setiap kelas melampaui nilai yang didapatkan dari proses pelatihan data. Dimana nilai matrix evaluasi pada saat pengujian langsung di lapangan menghasilkan nilai precision mencapai 100%, recall 88,8%, dan F1 Measure 94,1%. Sistem secara keseluruhan dapat memberikan peringatan dini dengan tingkat akurasi yang baik untuk meningkatkan keselamatan nelayan.
======================================================================================================================================
At the end of 2021, the Komite Nasional Keselamatan Transportasi (KNKT) recorded that around 31 percent of ship accidents between 2018 and 2020 occurred on fishing vessels. The KNKT stated that around 100 people per year died in these fishing vessel accidents. According to KKP findings, most traditional fishing vessels are not equipped with navigation tools, leading to frequent collisions with other vessels, including both large ships and other fishing vessels. Therefore, this thesis aims to develop an automatic vessel detection system to enhance safety during fishing vessel operations. To develop the automatic vessel detection system, studies on collision avoidance, object detection systems, and machine learning are required. In this thesis, the YOLOv8 algorithm was utilized to train the obstacle image database obtained with the assistance of the Google Image Scraper. There are four classifications of obstacles detected, namely fishing boats, tugboats, dredging vessel, and cargo ships. The hardware design was carried out using equipment such as a Raspberry Pi, serving as a mini PC and the brain of the automated obstacle detection system and early-warning collision detection system on ships. The best model was obtained after 120 iterations, with the evaluation matrix values for each class exceeding those obtained during the data training process. During field testing, the evaluation matrix values achieved a precision of 100%, a recall of 88.8%, and an F1 Measure of 94.1%. The system as a whole can provide early warnings with reasonable accuracy, thereby enhancing the safety of fishermen.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Penghindaran Tabrakan, Obstacle detection, Kapal ikan, Computer Vision, Navigasi Maritim. Collision Avoidance, Obstacle detection, Fishing Vessel, Computer Vision, Maritime Navigation.
Subjects:	T Technology > T Technology (General) > T58.62 Decision support systems V Naval Science > VK > VK555 Navigation. V Naval Science > VM Naval architecture. Shipbuilding. Marine engineering V Naval Science > VM431 Fishing boats V Naval Science > VM Naval architecture. Shipbuilding. Marine engineering > VM471 Ships--Electric equipment
Divisions:	Faculty of Marine Technology (MARTECH) > Naval Architecture and Shipbuilding Engineering > 36201-(S1) Undergraduate Thesis
Depositing User:	Novalyina Ghassani Putri
Date Deposited:	07 Aug 2025 01:01
Last Modified:	07 Aug 2025 01:01
URI:	http://repository.its.ac.id/id/eprint/126309

Actions (login required)

View Item