基于计算机视觉的船模位置跟踪  

Position tracking of model ships based on computer vision

作　　者：陈增志[1] 白响恩[1] 吴善刚[1] 肖英杰[1] 陈亮[1] 

机构地区：[1]上海海事大学商船学院,上海201306

出　　处：《上海海事大学学报》2013年第3期27-31,共5页Journal of Shanghai Maritime University

摘　　要：针对水池实验室中船模位置的跟踪问题,提出计算机视觉跟踪的方法.首先对视频图像进行逆透视投影矫正,然后利用CamShift算法实时跟踪船模在图像中的重心位置,最后将船模的图像坐标换算成实际空间坐标,实现船模位置的精确跟踪.该方法具有系统工作稳定性高,安装简单方便,跟踪精度高、效果较好等特点,同时还可利用跟踪得到的位置坐标,计算出船模的航速和航迹向,以实现这两个状态参数的跟踪,为实时跟踪船模等课题研究打下基础.About position tracking of model ships in water tank laboratories, a tracking method based on computer vision is proposed. First, video images are corrected by an inverse perspective projection meth- od. Then, a CamShift algorithm is used to real-timely track center-of-gravity positions of model ships in images. At last, the coordinates of model ships in images are transformed into real spatial coordinates to achieve precise position tracking of model ships. The method has good characteristics, such as high sta- bility system, easy installation, and high-accuracy and better-effect tracking. At the same time, the posi- tion coordinates of model ships obtained by tracking can be used to calculate velocities and directions of model ships. Thus, a foundation can be laid for achieving these two state parameter tracking and relevant study on real-time tracking of model ships.

关 键 词：船模跟踪 CAMSHIFT算法 计算机视觉 OPENCV 

分 类 号：U666.158[交通运输工程—船舶及航道工程] U675.6[交通运输工程—船舶与海洋工程]