【作者】 王正言；

【导师】 刘俊； 杨卫；

【作者基本信息】 中北大学 ， 测试计量技术及仪器， 2012， 硕士

【摘要】 目标跟踪技术无论在军用还是民用方面都具有非常重要的意义，是当今世界各国学者研究的热点和前沿问题之一。本文进行了红外图像目标跟踪技术的硬件及软件系统研究。本文以静止背景下的人为研究对象，其研究目的是在非人为干预的情况下，应用图像处理、智能分析技术等对捕获的图像进行理解、分析。以完成对应用场景中运动目标的检测及跟踪等任务。根据项目的发展需求，进行多路显示技术的研究，旨在扩大监测范围，提供更全面的监测信息。首先，本文对图像目标跟踪及多路显示硬件电路技术进行研究，设计基于DSP的目标跟踪及多路显示硬件系统。采用TI公司专为视频、图像处理及应用，而设计的DSP微处理器芯片TMS320DM642(以下简称DM642)。研究、设计该芯片及其外围电路组成的硬件平台。论文进行了系统方案论证、硬件原理图设计以及测试工作等。介绍了硬件电路里各个模块的设计，包括图像采集、时序控制、图像编码输出等模块；最后对系统进行测试，得到较理想效果。其次，研究图像的预处理算法，应用相应的方法来弥补在视频图像采集时由于应用场景、光照、机器本身等产生的噪声等干扰。运用基于三帧差法的目标检测算法，来实现运动目标的检测。利用形心跟踪的方式完成对目标的跟踪任务，编写相关软件并进行实验得到较理想的跟踪效果。最后，本文旨在将多路CCD捕获的场景信息同时显示在一个监视画面上，通过对图像缩放算法的研究、多路显示框架的研究，实现多路视频图像信息同时、实时的显示。更多还原

【Abstract】 Whether in military or civilian, target tracking technology has very importantsignificance, which is hot and one of the cutting-edge issues of the countries in the world ofacademic researchers’. This paper mainly studies the hardware and software systems of theinfrared image target tracking technology. The human of stationary background is researchsubject，in this paper. Which research purpose is under the case of non-human intervention, tounderstand and analyze images captured by the camera of the application of imageprocessing, video processing technology and intelligent analysis technology. To completethe task this is detection and tracking of moving targets in the Scenarios. Considered tocomply with project requirements, conduce multi-channel display technology, aimed atexpanding the scope of monitoring, providing more comprehensive monitoring information.Firstly, the paper research on image tracking and multi-channel display hardware circuittechnology, and design the target tracking and multi-channel display based on DSP hardwaresystem. Research and design the hardware platform that the main chip and itsperipheral circuit. In this paper using the DSP chip on TMS320DM642 (Hereinafter referredto as DM642) which is TI Company’s designed for video, image processing and application.Than do the follow job conducted demonstration of the program, the design of the hardwareschematic，layout design and testing work and so on. Introduced the design of eachmodule in the hardware circuit , including image acquisition module, timing controlmodule, image coding output module and so on; Finally, it gets more satisfactory results aftertesting the system.Secondly, the paper study the image preprocessing algorithm, design the appropriate wayto compensate for the noise in the video image acquisition due to the scenarios, light, machineitself. The use of target detection algorithm based on three difference method, which achievesthe moving target detection. Using the cancroids tracking methods complete the target tracking task, compile related software and experiment with better tracking.Finally，This paper aims to make the scene information which captured by multi-channelCCD display on a monitor screen, through researching the image scaling algorithm and themulti-channel display framework to achieve multi-channel video image information at thesame time, real-time display.更多还原