图像理解与机器视觉论文.docVIP

图像理解与机器视觉报告 课程编号：0441012 课程名称：《图像理解与机器视觉》 课程学分： 3.0 学生姓名：高黎 学号：1076490489 学科专业：控制工程 导师：陈光达 机器视觉动态目标跟踪 高黎 Abstract This dissertation consists some parts. The major work dissertation include: 1. Introduce the research of active machine vision, the important of the project which introduced the attention select and object tracking. Through this part, we can understand the structure and the research of the project. 2．In order to understand active machine vision in more detail, we studyed the principle of active machine vision, analysis the mechanical and information transmission model of the system. Designed system structure, analysised its character .on the base of principles, we list some kinds of camera heads, given the design at last. 3. Images are input of binocular vision system when research object tracking and coordination, many image dispose algorithms are used. Stereo vision method is used in fixation, to get spacial positions .so this part analysised image dispose and stereo vision methods. 4. Research tracking and algorithm of coordination of multi-DOFsystem, simulations and experiments were made to verify the algorithms. Finally,the paper point out the problems and propose the fields of the future researches. The first, the real time tracking is influenced by delay of mechanical factors and speed of image dispose; the second, noise should consider in further research; the third, some problem as one eye of cameras be blinded should be considerd. 摘要 本文的主要工作有以下几个部分： 1．讨论了主动机器视觉系统的原理，分析系统的机理模型和信息处理模型，并设计系统的硬件结构。在系统原理的基础之上，给出了系统的功能模块。 2．图像是双目视觉系统的输入信息，研究动态目标跟踪和多自由度协调控制的时候，要用到很多图像处理算法和立体视觉方法，本文介绍了有关的图像处理方法和立体视觉方法。 3．主动机器视觉系统的基本功能是完成注意力选择、定位、跟踪。跟踪是个重要的部分。目标的跟踪是主动机器视觉的任务之一，是为上层认知与识别层提供所需信息的途径。 4. 进行了主动机器视觉系统动态目标跟踪和多自由度协调控制实验和仿真研究。由于这个系统涉及到图像处理、立体定位、标定等问题，甚至还有高层的图像描述问题，所以这里先主要研究系统结构中的参数（基线长度）B的选取；然后对图像处理算法和多自由度协调控制算法进行试验和仿真，分析了实验和仿真结果。 虽然对主动机器视觉动态目标跟踪的研究取得了一些成果，但是还有一些问题需要进一步的研究。首先，研究中没有对系统的实时性进行分析。由于