Researchers at Indonesia's IPB University are developing a system of robots and artificial intelligence to monitor marine ecosystems, while another university and a conservation group have adapted an app to record fish catches. The app's AI system has catalogued about 25 species.
Indra Jaya, a professor at IPB's Faculty of Marine Affairs and Fisheries, leads the work combining surface and underwater observation vehicles with AI. A surface drone collects data on aquatic conditions, while underwater vehicles inspect floating fish cages, coral density and the abundance of marine life.
Indra said AI helps researchers identify and classify what they observe and make predictions. He said a complete prototype had been developed, with field trials through a pilot project planned before any large-scale rollout in Indonesia.
In Sumedang, West Java, Yayasan Konservasi Alam Nusantara (YKAN) and researchers at Universitas Padjadjaran (UNPAD) have adapted the Crew-Operated Data Recording System, or CODRS, into an Android app that works offline. Fishers photograph their catch on a standard measuring board, and AI identifies the species and estimates its size. The photos and catch coordinates are uploaded for verification when the vessel reconnects to the internet.
UNPAD AI developer Aurelia Norma said the data could help track whether species remain in an area and whether overfishing may be occurring. The system has recorded the morphology and size indicators of about 25 species, with verification accuracy above 80 percent. UNPAD marine conservation student Fenyawati Akhmed said more precise records could help monitor fish stocks.
Syahril Abd. Raup, director of fisheries resources management at Indonesia's Ministry of Marine Affairs and Fisheries, said an AI module that identifies fish species and size had been integrated into E-Logbook Version 3. It supports the country's quota-based measured fishing policy. The data also provides a basis for planning the management of marine protected areas. Reporting was supported by Internews-Earth Journalism Network.