Satellite Detection of Moving Vessels in Marine Environments

Abstract

There is a growing need for coverage of large maritime areas, mainly in the exclusive economic zone (EEZ). Due to the difficulty of accessing such large areas, the use of satellite based sensors is the most efficient and cost-effective way to perform this task. Vessel behavior prediction is a necessary ability for detection of moving vessels with satellite imagery. In this paper we present an algorithm for selection of the best satellite observation window to detect a moving object. First, we describe a model for vessel behavior prediction and compare its performance to two base models. We use real marine traffic data (AIS) to compare their ability to predict vessel behavior in a time frame of between 1–24 hours. Then, we present a KINGFISHER, maritime intelligence system which uses our algorithm to track suspected vessels with satellite sensor. We also present the results of the algorithm in operational scenarios of the KINGFISHER.

A Foraging Strategy with Risk Response for Individual Robots in Adversarial Environments

“…In this section, we study how the robot performs adversarial foraging tasks in adversarial environments, where both the resources utilities and the risk levels are assumed homogeneous. A motivating example is in adversarial information-gathering applications, where a UAV is required to collect information about opponents on the battleields; the risks in this adversarial environment exist in the form of enemy agent patrolling [28,48]. Many studies in multiagent adversarial patrolling domains generally assume that patrolling points often have the same utilities, and conclude that it is optimal for patrolling agents to visit these patrolling points with the same probabilities [2,65].…”

“…If a UAV is required to collect information in this adversarial environment, the utilities gained by sending any collected information back are the same. Therefore, the resource utilities in the environment are assumed homogeneous; • Hypothesis on the homogeneous risk levels: Based on the above hypothesis on homogeneous resource utilities, many studies have pointed out that it is optimal for patrolling agents to visit these patrol points of equal utility with the same frequency according to Nash equilibrium [2,28]. The same frequency leads to the same probability of the robot being detected on each edge.…”

“…From this, it is shown that in many studies, the resource utilities are usually assumed homogeneous. Based on the hypothesis on homogeneous resource utilities, Agmon et al [2,28] concluded that homogeneous risk levels result in an optimal adversarial strategy. Without loss of generality, the entire environment is also assumed homogeneous in this section, where all regions are associated with the same risk level (i.e., the survival probability p i j is equal for each edge e i j ∈ E), and the objects available at each foraging location have the same utility (i.e., the utility д i is also equal for each l i ∈ L t ).…”

Encoding Temporal and Spatial Vessel Context using Self-Supervised Learning Model (Student Abstract)

“…Other related works model vessel behavior from historical AIS information using probabilistic models, e.g. Auslander, Gupta, and Aha (2012) and Fridman et al (2019) construct Markov models, such as Markov logic networks and Markov chains, or hierarchical neural networks (Kim and Lee 2018). Nguyen et al (2018Nguyen et al ( , 2019 have explored the idea of multitask machine learning.…”

Automatic Ship Recognition Chain on Satellite Multispectral Imagery

“…In this context, the imagery is used to precisely track and monitor each ship by maritime authorities. As a result, by combining fine quality imagery and a reliable developed detection algorithm, suspicious activities such as illegal fishing can be identified [8] - [12]. Another technology that can also be implemented for supporting maritime awareness service is the Automatic Identification System (AIS).…”