基于固定/PTZ摄像机系统的开阔水域小目标船舶主动式跟踪方法  

An Active Tracking Method for Small Ships in Open Water Based on Fixed/PTZ Camera System

作　　者：游继安 胡钊政[1,3,4] 肖汉彪 孟杰[1,3,4] YOU Ji'an;HU Zhaozheng;XIAO Hanbiao;MENG Jie(State Key Laboratory of Maritime Technology and Safety,Wuhan University of Technology,Wuhan 430063,China;College of Technology,Hubei Engineering University,Xiaogan 432000,China;Hubei Key Laboratory of Transportation Internet of Things,Wuhan University of Technology,Wuhan 430063,China;Intelligent Transportation Systems Research Center,Wuhan University of Technology,Wuhan 430063,China)

机构地区：[1]武汉理工大学水路交通控制全国重点实验室,武汉430063 [2]湖北工程学院新技术学院,湖北孝感432000 [3]武汉理工大学交通物联网技术湖北省重点实验室,武汉430063 [4]武汉理工大学智能交通系统研究中心,武汉430063

出　　处：《交通信息与安全》2024年第3期53-61,共9页Journal of Transport Information and Safety

基　　金：国家自然科学基金项目(52302503);湖北省重点研发计划项目(2022BAA064);湖北省教育厅科学研究计划指导性项目(B2022510)资助。

摘　　要：仅依靠当前的闭路电视(closed-circuit television,CCTV)系统,往往难以主动跟踪并拍摄内河船舶清晰的图像。针对上述问题,研究了基于固定/平移-倾斜-对焦(pan-tilt-zoom,PTZ)摄像机系统的开阔水域小目标船舶主动式跟踪方法。采用基于虚拟四边形的三层联合标定模型对固定摄像机和PTZ摄像机进行联合标定,将虚拟四边形内的图像坐标与PTZ摄像机的平移角和倾斜角一一对应;引入虚拟四边形的概念,有效过滤虚拟四边形外目标的干扰,提高小目标船舶的检测准确率;利用透视n点(perspective-n-point, PnP)问题算法和虚拟四边形顶点的图像坐标,得到图像坐标与世界坐标间的映射关系,再利用Pan-Tilt-Height(PTH)模型将虚拟四边形中目标的世界坐标转化为PTH坐标;在小目标跟踪过程中,通过连续检测虚拟四边形中船舶边框质心的图像坐标,即可计算得到PTZ摄像机的平移角与倾斜角,从而实现实时主动跟踪的目的,并尽最大限度的保持船舶目标处于PTZ摄像机图像的中心位置。选取湖北省孝感市春晖湖和武汉市汉江中法桥段这2处真实场景,进行可靠性和有效性验证,实验结果表明:①利用改进的目标检测方法对固定摄像机图像中的船舶进行检测,F1-Score分别为96.82%和97.62%;②利用研究的主动式跟踪方法跟踪运动船舶时,PTZ摄像机的跟踪失败率为4.63%。本文研究的主动式跟踪方法的跟踪速率可以达到18.55 fps。It is difficult to actively track and capture clear images of inland river ships with the current Closed Circuit Television(CCTV)system.To fill the gap,an active tracking method for small ships in open waters based on the fixed/pan-tilt-zoom(PTZ)camera system is proposed.A three-layer joint calibration model based on a virtual quadrilateral(VQ)is introduced to jointly calibrate the fixed camera and the PTZ camera,which matches the image coordinate with the pan and tilt angle of the PTZ camera one by one;The introduced VQ filters out the targets outside the quadrilateral,eliminating inference and improving the accuracy of detection.The mapping relationship between the image coordinates and the world coordinates can be obtained by using the Perspective-n-Point(PnP)algorithm and the vertices of the VQ;Fourthly,the world coordinates of the points in the VQ are transformed into the Pan-Tilt-Hight(PTH)coordinates via PTH model.Then,by calculating the coordinate of the ship(the centroid of the ship)in the VQ,the pan and tilt angle of the PTZ camera can be derived,achieving real-time active tracking and keeping the target at the center of the PTZ camera image.To validate the proposed method,two real scenes are introduced,namely Chunhui Lake in Xiaogan City and the Sino-French Bridge section of the Han River in Wuhan City,Hubei Province.The results indicate that,①the F1-Scores of the proposed method on the fixed camera are 96.82%and 97.62%,respectively;②when the proposed method is applied to the PTZ camera for tracking the moving ships,the failure rate is 4.63%.In summary,the proposed active tracking method performs reasonably in practice with a high tracking rate of 18.55 fps.

关 键 词：智能交通 小目标船舶 主动式跟踪 联合标定 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]