ARTIFICIAL INTELLIGENCE FOR MARITIME SAFETY AND ENVIRONMENTAL SURVEILLANCE IN INDONESIA: AN OFFICIAL-STATISTICS BASELINE

Ifan Widayatullah; Muhammad Ikhsan Setiawan; Mohd Haziman Wan Ibrahim

doi:10.29138/neutron.v24i01.298

Ifan Widayatullah Narotama University
Muhammad Ikhsan Setiawan Maritime University AMNI Semarang
Mohd Haziman Wan Ibrahim Universiti Tun Hussein Onn

DOI: https://doi.org/10.29138/neutron.v24i01.298

Keywords: Artificial Intelligence, Indonesia, K-Means Clustering, Maritime Transport, Official Statistics, Oil-Spill Detection, Vessel Monitoring

Abstract

Maritime safety and environmental surveillance in Indonesia depend on the reliability of official data supporting artificial intelligence (AI). This study establishes an official-statistics baseline using BPS-Statistics Indonesia’s Marine and Coastal Resources Statistics 2023. It examines fishing-vessel fleets for 2021–2022 (34 provinces; 1,004,060 and 893,258 vessels), 118 classified fishing ports, 41 recorded oil-spill incidents (1999–2023), port-water monitoring at 27 sites and fisheries patrol records. Methods include concentration indices, k-means clustering with silhouette selection, principal component analysis, modified z-scores and Isolation Forest screening. In 2022, 88.8% of the recorded fleet was motorised; provincial distribution showed a Gini coefficient of 0.41. Three fleet typologies were identified: outboard-dominated, inboard-dominated and non-motorised-heavy. Ten of 94 provincial vessel-type changes were flagged as statistical outliers, while 10 of 27 water-monitoring sites reported none of three core parameters. In 2021, there was approximately one recorded inspection per 385 fishing vessels. These descriptive findings inform candidate AIS trajectory analytics, SAR oil-spill detection, optical satellite monitoring and smart-port planning. The study proposes a phased, data-first roadmap; it does not estimate the performance of operational AI systems in Indonesian waters.