EUVIMEDTutustu EUVIMEDiin
Crossref

Deteksi Kapal Menggunakan Algoritma DC-YOLOV8

Freddy Winarso, Arief Setyanto, Anggit Dwi Hartanto

Lambda: Jurnal Ilmiah Pendidikan MIPA dan Aplikasinya · 2026-02-06 · 5 (4) · 1115-1124

DOI: 10.58218/lambda.v5i4.1926 ↗

At a glance

Journal article in Lambda: Jurnal Ilmiah Pendidikan MIPA dan Aplikasinya, published online on 2026-02-06. Written by Freddy Winarso and 2 co-authors.

TiivistelmäBibliografiset tiedotCiteRelated studies

Abstract (original language)

This study aims to improve the inference time and accuracy of ship detection in complex maritime imagery using the DC-YOLOv8 algorithm. DC-YOLOv8 is an enhanced version of YOLOv8 specifically designed for real-time detection under challenging visual conditions, such as low lighting, fog, and dynamic sea backgrounds. The method integrates an MDC module, depthwise separable convolution, and an updated feature fusion strategy. The evaluation was conducted using the SeaShips dataset and compared with the baseline YOLOv8 model. Experimental results show that DC-YOLOv8 achieves a significant improvement in inference time, reducing it by 2.855 ms, while maintaining a competitive mAP score of 0.7953. These results indicate that DC-YOLOv8 is more effective for real-time detection tasks in maritime environments.

Cite

APA

Winarso, F., Setyanto, A., & Hartanto, A. D. (2026). Deteksi Kapal Menggunakan Algoritma DC-YOLOV8. Lambda: Jurnal Ilmiah Pendidikan MIPA dan Aplikasinya, 5(4), 1115-1124. https://doi.org/10.58218/lambda.v5i4.1926

Vancouver

Winarso F, Setyanto A, Hartanto AD. Deteksi Kapal Menggunakan Algoritma DC-YOLOV8. Lambda: Jurnal Ilmiah Pendidikan MIPA dan Aplikasinya. 2026;5(4):1115-1124. doi:10.58218/lambda.v5i4.1926

BibTeX
@article{Winarso2026,
  author = {Winarso, Freddy and Setyanto, Arief and Hartanto, Anggit Dwi},
  title = {Deteksi Kapal Menggunakan Algoritma DC-YOLOV8},
  journal = {Lambda: Jurnal Ilmiah Pendidikan MIPA dan Aplikasinya},
  year = {2026},
  volume = {5},
  number = {4},
  pages = {1115--1124},
  doi = {10.58218/lambda.v5i4.1926}
}
Start for free →