本文選題：仿生眼　切入點(diǎn)：眼球運(yùn)動(dòng)控制　出處：《山東大學(xué)》2016年碩士論文

【摘要】：人的眼睛是感知外界事物,與環(huán)境進(jìn)行交互的重要器官,在其對(duì)外界目標(biāo)事物進(jìn)行感知、識(shí)別、跟蹤等過程中具有復(fù)雜而巧妙的神經(jīng)控制機(jī)制,能幫助人類快速、準(zhǔn)確地對(duì)環(huán)境或者是目標(biāo)事物進(jìn)行認(rèn)知。機(jī)器視覺的快速發(fā)展使機(jī)器具備與外界環(huán)境進(jìn)行交互的基本能力,但是與人眼的高級(jí)視覺功能的差距還是很大。因此,深入理解人類眼球的控制機(jī)理,將其與機(jī)器視覺和控制科學(xué)等技術(shù)進(jìn)行融合創(chuàng)新,提高機(jī)器的認(rèn)知功能,具有重要的研究意義。本文重點(diǎn)對(duì)仿生眼運(yùn)動(dòng)控制算法進(jìn)行了研究,主要工作如下：首先,簡(jiǎn)單分析了仿生眼研究的背景與意義,對(duì)當(dāng)前仿生眼的研究現(xiàn)狀進(jìn)行了回顧與總結(jié),指出了現(xiàn)階段仿生眼研究存在的主要問題,介紹了本文的主要研究?jī)?nèi)容和后續(xù)的章節(jié)安排。第二,研究了人眼球的結(jié)構(gòu)特點(diǎn),分析了其獲取圖像的原理。對(duì)眼球運(yùn)動(dòng)控制的神經(jīng)機(jī)制進(jìn)行了分析,為搭建仿生眼控制模型打下了生理學(xué)的基礎(chǔ)；探討了仿生眼運(yùn)動(dòng)控制與仿生視覺感知、理解的關(guān)系。第三,深入分析了眼球平滑追蹤運(yùn)動(dòng)的神經(jīng)控制機(jī)制,根據(jù)神經(jīng)控制機(jī)理建立了基于卡爾曼濾波和迭代學(xué)習(xí)控制算法的平滑追蹤模型。眼球獲取的圖像通常含有噪聲,卡爾曼濾波算法可以濾除這種干擾,保證了輸入的有效；迭代學(xué)習(xí)控制可以有效的解決周期性的目標(biāo)跟蹤存在的問題,通過對(duì)前一周期目標(biāo)運(yùn)動(dòng)信息的存儲(chǔ)學(xué)習(xí),可以得到當(dāng)前周期目標(biāo)運(yùn)動(dòng)最優(yōu)的控制信號(hào),實(shí)現(xiàn)對(duì)目標(biāo)的完全跟蹤,解決了因處理圖像所帶來的時(shí)延問題；通過參數(shù)整定的方法得到了最優(yōu)的迭代學(xué)習(xí)控制參數(shù),增加了模型的自適應(yīng)性；通過對(duì)比仿真實(shí)驗(yàn)對(duì)算法進(jìn)行了驗(yàn)證。第四,搭建仿生眼硬件實(shí)驗(yàn)平臺(tái)。介紹了仿生眼硬件平臺(tái),設(shè)計(jì)出仿生眼驅(qū)動(dòng)電路,并利用箱體結(jié)構(gòu)進(jìn)行集成改進(jìn)仿生眼平臺(tái)結(jié)構(gòu)；采用PTZ攝像頭作為仿生眼研究的實(shí)驗(yàn)平臺(tái),并對(duì)簡(jiǎn)單的眼球跟蹤算法進(jìn)行了實(shí)驗(yàn)驗(yàn)證；為了實(shí)現(xiàn)對(duì)目標(biāo)運(yùn)動(dòng)狀態(tài)的控制,設(shè)計(jì)了滑軌運(yùn)動(dòng)平臺(tái),在滑軌平臺(tái)和PTZ攝像頭的基礎(chǔ)上對(duì)基于迭代學(xué)習(xí)控制算法的平滑追蹤模型進(jìn)行了實(shí)驗(yàn)驗(yàn)證。最后,對(duì)本文的研究?jī)?nèi)容進(jìn)行了總結(jié),并對(duì)未來的工作方向進(jìn)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.
【學(xué)位授予單位】：山東大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2016
【分類號(hào)】：TP391.41

【相似文獻(xiàn)】

相關(guān)重要報(bào)紙文章 前1條

1 邵君;夏秋季更要保護(hù)眼睛[N];醫(yī)藥養(yǎng)生保健報(bào);2009年

相關(guān)碩士學(xué)位論文 前5條

1 李澤東;年齡相關(guān)性黃斑變性微掃視眼球運(yùn)動(dòng)特征的臨床研究[D];天津醫(yī)科大學(xué);2015年

2 秦斌;面向人類眼球運(yùn)動(dòng)機(jī)理的仿生眼運(yùn)動(dòng)控制算法研究[D];山東大學(xué);2016年

3 王強(qiáng);眼球水平運(yùn)動(dòng)的生物動(dòng)力學(xué)多元時(shí)序模型的建立及其研究[D];山東大學(xué);2006年

4 李玉秀;基于IPMC驅(qū)動(dòng)的仿生眼球設(shè)計(jì)及制造[D];南京航空航天大學(xué);2011年

5 張李娜;鞏膜搜索線圈對(duì)微掃視性眼球運(yùn)動(dòng)幅度影響的觀察與分析[D];天津醫(yī)科大學(xué);2014年

，

本文編號(hào)：1672965