关键词: 船舶自动识别系统, 云台摄像机, 控制河段, 预测跟踪, 视觉伺服

Abstract: It is difficult to effectually track and monitor a vessel by video monitoring in narrow and winding controlled areas of Yangtze River. In this situation, directly using the common method for maritime video surveillance would result in tracking lag, alignment error, and even loss of target. A method of combining Automatic Identification System (AIS) and multiple Pan-Tilt-Zoom (PTZ) cameras was proposed to trace and monitor the vessel in controlled areas of Yangtze River. The latitude and longitude of the target vessel were utilized to initialize the alignment of PTZ cameras. Meanwhile, vessel’s speed and heading were applied for dynamic prediction tracking to smoothen image and eliminate tracking lag, while image recognition was employed to correct error for ensuring that target vessel was located in the optimum surveillance area.The simulation results indicate that this algorithm can not only track target vessel continuously and dynamically, but also realize the real-time target tracking with high precision.

Key words: AIS, PTZ camera, controlled areas, prediction tracking, visual servo