Tracking fishing vessels plays a pivotal role in fisheries monitoring, control, and surveillance. The economic impact from illegal, unreported, and unregulated fishing ranges between $10 and $23 billion USD globally. This has attracted many researchers' focus towards addressing the problem of illegal fishing. Fishing trip is the most appropriate granularity level to study routine fishing activity patterns to detect suspicious activities. Since self-reported information about fishing vessel trips is notoriously unreliable, we propose here an unsupervised learning approach to partition raw trajectories of vessels engaging in fishing into trips and identify trip types. Our approach, first partitions a fishing trip into micro-activities, then uses cluster analysis to confirm the micro-activity type. Next, it employs multiple Hidden Markov Models to partition the trip into segments, each of which represents a routine activity. Finally, our proposed method utilizes maritime contextual information to differentiate various fishing trip types, revealing actionable knowledge about vessel activities and their operations. Our experimental evaluation on a large real-world fishing vessel trajectory dataset confirms our method's practicability and effectiveness for enhancing maritime domain awareness.
Copyright is held by the author(s).
This thesis may be printed or downloaded for non-commercial research and scholarly purposes.
Supervisor or Senior Supervisor
Thesis advisor: Glässer, Uwe
Member of collection