Computing with Words in Maritime Piracy and Attack Detection Systems
Computing with Words in Maritime Piracy and Attack Detection Systems
复制标题
海上海盗和攻击检测系统中的文字计算
DOI：
10.1007/978-3-030-50439-7_30
复制
发表时间：
2020
期刊：
影响因子：
--
通讯作者：
A. Kandel
中科院分区：
文献类型：
--
作者：
Jelena Tešić;D. Tamir;Shai Neumann;N. Rishe;A. Kandel
.在本文中，我们建议将深度学习的最新进展应用于设计和训练算法，以在不同条件下（例如，暴风雪、强光、夜晚），以及在使用文字进行计算以识别威胁活动，其中缺乏训练数据阻止使用深度学习。最近，海盗和袭击运输船只的行为增加，给全球经济造成数十亿美元的损失。为了应对这一威胁，研究人员提出了代理驱动的建模来捕捉海上运输系统的动态，并对一系列海盗对策的潜力进行评分。将来自机载传感器和摄像头的信息与外部来源的情报相结合，用于早期海盗威胁检测，已显示出有希望的结果，但缺乏实时更新的情况。这种系统可以受益于早期预警，例如“一艘船正在接近该船并加速”、“一艘船正在绕船盘旋”或“两艘船正在偏离该船”。现有的机载摄像头捕捉这些活动，但没有自动处理程序的这种类型的模式，以通知预警系统。只有在机组人员收到可能发生攻击的警报后，他们才使用视觉数据馈送。摄像头传感器价格低廉，但将传入的视频数据流转换为可操作的项目仍然需要昂贵的人工处理。
. In this paper, we propose to apply recent advances in deep learning to design and train algorithms to localize, identify, and track small maritime objects under varying conditions (e.g., a snowstorm, high glare, night), and in compu-ting-with-words to identify threatening activities where lack of training data pre-cludes the use of deep learning. The recent rise of maritime piracy and attacks on transportation ships has cost the global economy several billion dollars. To counter the threat, researchers have proposed agent-driven modeling to capture the dynamics of the maritime transportation system, and to score the potential of a range of piracy countermeasures. Combining information from onboard sensors and cameras with intelligence from external sources for early piracy threat detection has shown promising results but lacks real-time updates for situational context. Such systems can benefit from early warnings, such as “a boat is approaching the ship and accelerating,” “a boat is circling the ship,” or “two boats are diverging close to the ship.” Existing onboard cameras capture these activities, but there are no automated processing procedures of this type of patterns to inform the early warning system. Visual data feed is used by crew only after they have been alerted of a possible attack. Camera sensors are inexpensive but transforming the incoming video data streams into actionable items still requires expensive human processing.