水下机器人作业目标控制和目标识别  

Autonomous Underwater Vehicle Job Controlling and Object Recognition

作　　者：陈荣盛[1] 袁小海[1] 胡震[1] 吴文伟[1] 刘培林 

机构地区：[1]中国船舶科学研究中心

出　　处：《船舶力学》1998年第1期62-68,共7页Journal of Ship Mechanics

摘　　要：自主式水下机器人（AUV）具有自主导航、自主避碰和自主作业能力，自主作业是水下机器人实现智能化的关键技犬。本文着重叙述AUV对特定作业一“水下目标探测与识别”作业控制和实现过程，叙述水下三维目标的识别机理和识别算法以及为正确识别目标提供最佳条件的预处理[1]、硬件系统和实时处理功能。最后给出试验结果，说明作业目标控制过程和识别算法具有实时性、高效率和可靠性。Autonomous Underwater Vehicle (AUV) can perform autonomous navigation, autonomous avoidance andautonomous job. Autonomous job is a key technique to realize AUV intelligence. The job controlling and implementation procedure for the special job-Underwater object detection and recognition are described in the paper. Thepaper includes underwater 3D object recognition mechanism, recognition algorithm, the optimal image pre-processingfor accurate recognition, hardware system and real-time processing. The experimental results given at last demonstrate that the job controlling and recognition algorithm has tile advantages in real-time,high efficiency and reliability.

关 键 词：水下机器人 作业控制 视觉信息 目标识别 

分 类 号：P754.3[天文地球—海洋科学] TP242[自动化与计算机技术—检测技术与自动化装置]