Object Identification, Classification and Avoidance in 3-D Underwater Automated Surveillance

3D 水下自动监视中的物体识别、分类和规避

• 批准号: 9506771
• 负责人: Kimon Valavanis
• 金额: $ 27.07万
• 依托单位: University of Louisiana at Lafayette
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 1995
• 资助国家: 美国
• 起止时间: 1995-09-01 至 1998-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9506771&HistoricalAwards=false
• 关键词: Object; Identification; Classification; Avoidance; Underwater

9506771 Valavanis The central mbjective of the proposed work is: object identification, classification and avoidance in 3-D underwater automated surveillance. The motivation behind this objective is monitoring and safeguarding wetlands, coastal fisheries and deep water polluted environments, which are among the most valuable national resources. The specific application domain, directly related to the proposed work, is surveying and monitoring Louisiana's wetlands and shallow water fisheries in collaboration with the Southern Science Center (SSC) - previously known as the National Wetlands Research Center - and the recently established National Marine Fisheries Services (NMFS), both located at the University of Southwestern Louisiana's (USL's) research park. This proposal capitalizes on already available resources, technology and algorithms, developed for other applications, but adapts and further modifies and enhances them for the proposed automated surveillance problem. The end result will be: a detailed methodology/approach for the real-time sensor based autonomous control and navigation of an underwater vehicle, capable of automated surveillance in sensitive environments. There will be no field studies and/or vehicle related experiments at this stage. Three specific objectives will be accomplished: i)Development of a detailed architecture for underwater vehicle motion control. ii) Implementation and experimental verification of an already derived total color difference (TCD) measure and total color difference tolerance (TCDT) thresholds for underwater object identification. iii) Enhancement and verification of a recently derived potential based panel method for autonomous vehicle motion planning in a 3-D workshop environment. ***

提出的工作的中心目标是：三维水下自动监视中的目标识别、分类和回避。这一目标背后的动机是监测和保护湿地、沿海渔业和深水污染环境，这些都是最宝贵的国家资源。与提议的工作直接相关的具体应用领域是与南方科学中心（SSC）（以前称为国家湿地研究中心）和最近成立的国家海洋渔业服务（NMFS）合作，对路易斯安那州的湿地和浅水渔业进行调查和监测，这两个中心都位于西南路易斯安那大学（USL）的研究园区。该方案利用了现有的资源、技术和算法，为其他应用开发，但针对所提出的自动化监控问题进行了调整、进一步修改和增强。最终结果将是：一种基于实时传感器的水下航行器自主控制和导航的详细方法/方法，能够在敏感环境中进行自动监视。在这个阶段将没有实地研究和/或车辆相关的实验。将完成三个具体目标：i)开发水下航行器运动控制的详细架构。ii)实现和实验验证已经导出的用于水下目标识别的总色差（TCD）测量和总色差容限（TCDT）阈值。iii)在三维车间环境中，增强并验证了一种基于电位的面板自动驾驶汽车运动规划方法。* * *