Rancang Bangun Sistem Identifikasi Jenis Dan Ukuran Ikan Berbasis Pattern Recognition Pada Kapal Penangkap Ikan Pelagis

Pranata, Ingwie Valentino Boy (2024) Rancang Bangun Sistem Identifikasi Jenis Dan Ukuran Ikan Berbasis Pattern Recognition Pada Kapal Penangkap Ikan Pelagis. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 04111940000011-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 April 2026. Download (4MB) | Request a copy

Abstract

Kementrian Kelautan dan Perikanan (KKP) menerapkan program penangkapan ikan terukur pada bulan Agustus 2022. Penangkapan ikan terukur yang dilakukan KKP di Pelabuhan Perikanan Nasional belum sepenuhnya bisa mengatasi krisis unreported fishing dan overfishing khususnya ikan pelagis karena kegiatan tangkap di laut lepas tidak terdeteksi secara real time seperti transhipment dengan pihak asing atau kapal berbendera asing. Sehingga tujuan dari penelitian ini adalah membuat sebuah sistem di kapal penangkap ikan pelagis untuk mengidentifikasi jenis dan ukuran ikan secara real time dan mempermudah proses tracking tangkapan ikan. Sampel atau dataset ikan pada penelitian ini adalah cakalang untuk ikan pelagis besar dan mackerel untuk ikan pelagis kecil. Penentuan dataset tersebut karena kemudahan dalam pengumpulan data gambar. Dilakukan pelabelan dan anotasi pada dataset. Training model untuk pattern recognition pada penelitian ini menggunakan algoritma YOLO dan CNN sebagai metodenya. Training model menghasilkan file checkpoint sebagai acuan dalam identifikasi jenis ikan yang masuk ke dalam sistem. Mean average precision (mAP) model dipilih yang terbaik di nilai 88%. Untuk prediksi panjang ikan pelagis didapat dari selisih x1 dan x2 dikalikan dengan koefisien. Prototype dibuat untuk memvisualisasikan sistem identifikasi. Prototype divisualisasikan sebagai web app dan diatur sedemikian rupa menyerupai sistem masuk ikan pelagis melalui konveyor, dirancang menggunakan Flask. Prototype ini memiliki fitur deteksi menggunakan metode gambar, video, serta kamera. Hasil uji coba, dari 8 inputan dapat dideteksi seluruhnya dengan benar. Tingkat akurasi deteksi berkisar antara 85% hingga 95%. Diharapkan penelitian ini dapat membantu penangkapan terukur Kementrian Kelautan dan Perikanan.
=============================================================================================================================
The Ministry of Maritime Affairs and Fisheries (KKP) implemented a measured fishing program in August 2022. Measured fishing conducted by KKP at the National Fishing Port has not been able to fully overcome the crisis of unreported fishing and overfishing, especially pelagic fish because fishing activities on the high seas are not detected in real time such as transhipment with foreign parties or foreign-flagged vessels. So the purpose of this research is to create a system on pelagic fishing vessels to identify the type and size of fish in real time and facilitate the process of tracking fish catches. The fish samples or datasets in this study are skipjack for large pelagic fish and mackerel for small pelagic fish. The determination of the dataset is due to the ease of collecting image data. Labeling and annotation are done on the dataset. Training model for pattern recognition in this research uses YOLO algorithm and CNN as the method. The training model produces a checkpoint file as a reference in identifying the type of fish that enters the system. The mean average precision (mAP) of the model was chosen as the best at 88%. For pelagic fish length prediction, it is obtained from the difference between x1 and x2 multiplied by the coefficient. A prototype was created to visualize the identification system. The prototype is visualized as a web app and arranged in such a way as to resemble a pelagic fish entry system through a conveyor, designed using Flask. This prototype has detection features using image, video, and camera methods. Test results, from 8 inputs can be detected. The detection accuracy rate ranges from 85% to 95%. It is hoped that this research can help the Ministry of Marine Affairs and Fisheries' measurable capture.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Ikan Pelagis, Pattern Recognition, Web App, Pelagic Fish
Subjects:	Q Science > Q Science (General) > Q337.5 Pattern recognition systems V Naval Science > VM431 Fishing boats
Divisions:	Faculty of Marine Technology (MARTECH) > Naval Architecture and Shipbuilding Engineering > 36201-(S1) Undergraduate Thesis
Depositing User:	Ingwie Valentino Boy Pranata
Date Deposited:	19 Feb 2024 01:17
Last Modified:	19 Feb 2024 01:17
URI:	http://repository.its.ac.id/id/eprint/107386

Actions (login required)

View Item