中文摘要：

为改进国产投弃式温深仪（XBT）的下降速率公式,采用基于图像处理的方法测量XBT探头的运动参数。采用背景差分法获取差分图像,选择合适的阈值并对差分图像二值化,标记二值图像中的连通区域,将二值图像分割为几个互相独立的区域。通过计算分析各连通区域的特征集合来识别XBT探头。利用探头中轴线描述探头的运动姿态,将探头的长轴作为探头的初始中轴线,然后通过迭代优化的方法得到更加精确的结果。实验结果表明：该方法角度测量平均误差小于0.5°,能满足XBT探头姿态测量精度要求。目前已将基于该算法搭建的视觉测量系统应用于XBT投放实验中,取得了令人满意的结果。

英文摘要：

In order to improve domestic-made XBT fall rate equation（ FRE）,use method based on image processing to measure movement parameters of XBT probe. Difference image is obtained by background difference method,appropriate threshold is selected and binaryzation of difference image is carried out,connected region of binary image is labeled and binary image is divided into several independent regions. The XBT probe is recognized by calculating and analyzing characteristic set of each connected region. Motion posture of probe is described by central axis of probe,long axis of XBT probe used as initial central axis of probe,and iterative optimization method is used to get more precise results. The experimental results show that average error of angle measurement is less than 0. 5° and the algorithm satisfies precision requirement of posture measurement of XBT probe. Vision measurement system based on this algorithm has been applied in XBT deployment experiments and results are satisfied.