面向人类眼球运动机理的仿生眼运动控制算法研究

发布时间：2018-03-27 19:43

本文选题：仿生眼　切入点：眼球运动控制　出处：《山东大学》2016年硕士论文

【摘要】：人的眼睛是感知外界事物,与环境进行交互的重要器官,在其对外界目标事物进行感知、识别、跟踪等过程中具有复杂而巧妙的神经控制机制,能帮助人类快速、准确地对环境或者是目标事物进行认知。机器视觉的快速发展使机器具备与外界环境进行交互的基本能力,但是与人眼的高级视觉功能的差距还是很大。因此,深入理解人类眼球的控制机理,将其与机器视觉和控制科学等技术进行融合创新,提高机器的认知功能,具有重要的研究意义。本文重点对仿生眼运动控制算法进行了研究,主要工作如下：首先,简单分析了仿生眼研究的背景与意义,对当前仿生眼的研究现状进行了回顾与总结,指出了现阶段仿生眼研究存在的主要问题,介绍了本文的主要研究内容和后续的章节安排。第二,研究了人眼球的结构特点,分析了其获取图像的原理。对眼球运动控制的神经机制进行了分析,为搭建仿生眼控制模型打下了生理学的基础；探讨了仿生眼运动控制与仿生视觉感知、理解的关系。第三,深入分析了眼球平滑追踪运动的神经控制机制,根据神经控制机理建立了基于卡尔曼滤波和迭代学习控制算法的平滑追踪模型。眼球获取的图像通常含有噪声,卡尔曼滤波算法可以滤除这种干扰,保证了输入的有效；迭代学习控制可以有效的解决周期性的目标跟踪存在的问题,通过对前一周期目标运动信息的存储学习,可以得到当前周期目标运动最优的控制信号,实现对目标的完全跟踪,解决了因处理图像所带来的时延问题；通过参数整定的方法得到了最优的迭代学习控制参数,增加了模型的自适应性；通过对比仿真实验对算法进行了验证。第四,搭建仿生眼硬件实验平台。介绍了仿生眼硬件平台,设计出仿生眼驱动电路,并利用箱体结构进行集成改进仿生眼平台结构；采用PTZ摄像头作为仿生眼研究的实验平台,并对简单的眼球跟踪算法进行了实验验证；为了实现对目标运动状态的控制,设计了滑轨运动平台,在滑轨平台和PTZ摄像头的基础上对基于迭代学习控制算法的平滑追踪模型进行了实验验证。最后,对本文的研究内容进行了总结,并对未来的工作方向进行了展望。
[Abstract]:The human eye is an important organ for perceiving external things and interacting with environment. In the process of perception, recognition, tracking and so on, human eyes have complex and ingenious neural control mechanisms, which can help human beings quickly. The rapid development of machine vision enables the machine to have the basic ability to interact with the external environment, but there is still a big gap with the advanced visual function of the human eye. It is of great significance to deeply understand the control mechanism of human eyeball, merge it with machine vision and control science, and improve the cognitive function of machine. In this paper, we focus on the research of bionic eye motion control algorithm. The main work is as follows: firstly, the background and significance of bionic eye research are briefly analyzed, the current research situation of bionic eye is reviewed and summarized, and the main problems in bionic eye research are pointed out. The main contents of this paper and the following chapters are introduced. Secondly, the structure of human eyeball is studied, the principle of image acquisition is analyzed, and the neural mechanism of eye movement control is analyzed. It lays a physiological foundation for building a biomimetic eye control model, and discusses the relationship between bionic eye motion control and bionic visual perception and understanding. Thirdly, the neural control mechanism of eye smooth tracking movement is deeply analyzed. Based on neural control mechanism, a smooth tracking model based on Kalman filter and iterative learning control algorithm is established. Iterative learning control can effectively solve the problem of periodic target tracking. The optimal control signal of the current periodic target motion can be obtained by learning from the storage of moving information of the previous periodic target, and the complete tracking of the target can be realized. The problem of delay caused by image processing is solved, the optimal iterative learning control parameters are obtained by parameter tuning method, and the self-adaptability of the model is increased. The algorithm is verified by comparative simulation experiments. Build the bionic eye hardware experimental platform. Introduce the bionic eye hardware platform, design the bionic eye driving circuit, and use the box structure to integrate and improve the bionic eye platform structure; use the PTZ camera as the experimental platform for bionic eye research, In order to control the moving state of the target, a sliding track motion platform is designed. The smooth tracking model based on iterative learning control algorithm is verified experimentally on the basis of slide platform and PTZ camera. Finally, the research content of this paper is summarized, and the future work direction is prospected.
【学位授予单位】：山东大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP391.41

【相似文献】