【作者】 李智；

【导师】 王向军；

【作者基本信息】 天津大学 ， 测试计量技术及仪器， 2004， 硕士

【摘要】 电视跟踪雷达是一种被动式成像跟踪设备，能在比较复杂的背景下按照标准的电视制式，提取与分离视场内的目标，确定目标位置，估计目标运动趋势，实现对目标的实时跟踪。它具有跟踪精度高、跟踪状态平稳、抗干扰能力强，分辨率高和成本低等特点，在军事上受到普遍重视，并在武器系统中已获得广泛的应用。电视跟踪雷达系统主要由影像探测部分、图像处理部分和伺服机构等组成。实现系统的数字化和智能化，同时从图像中获取更多的目标运动参数信息，是电视跟踪技术的发展趋势。电视跟踪系统组成中的图像处理部分又称为跟踪处理器，是系统实现目标捕获、识别和目标运动参数估计的关键环节，系统的数字化和智能化设计主要指的就是这一部分的设计。光流法和角点跟踪法是两种重要的跟踪方法，在目标识别和运动参数估计方面有着重要的应用。实验表明，两种方法各有优点和改进的余地，在实验的基础上本文提出了特征角点跟踪法，这种方法对目标表达简洁，在目标识别和运动分析方面具有明显优势。本论文研究的题目是智能炮瞄电视跟踪雷达的数字化和信息获取方法。主要内容包括：数字化电视跟踪系统的总体设计思想和主要实现方法。讨论了电视跟踪系统设计的主要技术环节。光流算法在数字化电视跟踪系统中的应用。研究了光流算法原理，及其常见应用和缺陷。用MATLAB仿真实现了基于光流法的序列图像特征提取、运动目标区域分割、目标运动方向和速度计算。角点跟踪算法原理及其实现。研究通过提取运动目标的角点特征，获得目标的结构特征及速度特征的方法，实现目标的较精确识别和速度估计。通过仿真实验，对多种角点提取、匹配算法进行了比较研究，结果表明本文采用和改进的角点跟踪算法可以实现对目标的准确角点跟踪。特征角点跟踪算法的原理及其实现。提出了特征角点的概念，研究了基于特征角点的目标识别、瞄准点选择及速度估计方法。对序列图像进行了仿真实验，结果表明本文对特征角点的定义和提取方法是恰当的，跟踪算法稳定，角点不易丢失，有很好的应用前景。更多还原

【Abstract】 Television tracking system is a kind of passive image recognizing and tracking equipment. It can separate the targets form the field, certain their position, estimate their movement trend and realize the real-time tracking of the targets, at complicated background and according to standard television system type. It has the characteristics of high tracking accuracy, steady tracking state, strong anti-interference ability, good resolution, low cost and so on. It has acquired the extensive application in weapon system because of its great military value.The TV tracking system is made of three main parts, which include the image detection device, the image processing part and the servo system. To realize the its numeralization and intelligence, get more motion parameters of the moving targets form the serial images, are the development trends of the tracking system. The image process part of the system is also called as the tracking process device, is the system’s key component to acquire and recognize the target and estimate its motion parameters. The system’s digitization and intelligence design primarily is to point this part’s design. The two important tracking methods based on optical flow algorithm and corner detection having got the important application in the aspects of target identifying to estimate with motion parameters have been discussed in detail. The experiment expresses, two kinds of methods is each to have the advantage with the spare lot of the improvement, on the experience foundation a characteristic corner tracking methods had been presented. The characteristic corner tracking method describes the target brevity and has the obvious advantage in the aspects of target identifying with motion analyzing.In this thesis we investigate the digital target tracking and information acquisition methods of TV tracking system that equipped on the warship. The main contents includes:The digital TV tracking system’s total design concept and the primarily realized methods. The key components of the system’s design have been discussed.The theory of optical flow field and its applications in tracking system has been studied. The optical flow algorithm has been test by Matlab programs.The theory and realization of the corner tracking method. The method to acquire the construction and motion characteristic of the target whit high accuracy based on withdrawing the corner features of the target has been studied. Some methods <WP=5>of extracting corner features and matching the corners have been compared by experimentation.The theory and realization of the characteristic corner tracking method. The concept of characteristic corner has been presented. The method is a new way for building a simple target’s model and increase calculating speed to realization the efficiently accurate tracking.更多还原