基于ransac的视线跟踪关键技术研究-电子与通信工程专业论文.docxVIP

摘要 摘要 万方数据 万方数据 万方数据 万方数据 摘要 随着人机交互技术研究的兴起，关于眼视线跟踪的技术也日益成熟。眼视线 跟踪系统中关键的问题就是关于用户头动问题和精度问题。 本文首先阐述了眼视线跟踪技术的研究背景，介绍了侵入式和非侵入式眼动 跟踪技术的特点及其发展。接下来，分别介绍眼视线跟踪系统中的标定算法，精 度问题以及头动解决方法。随后，详细论述了眼视线跟踪关键技术，主要包括眼 图处理模块和标定模块的设计，着重论述眼图处理模块中，瞳孔参数的提取算法 的优化，使用的是基于 3 点的随机采样一致性（RANSAC）的瞳孔提取算法，并 从定位准确率方面详细分析了该算法的性能。 最后，对眼视线跟踪系统中鲁棒性标定方法进行了研究。首先，分析标定过 程中由于眨眼、视线游移等原因造成的干扰点对对系统性能的影响。随后提出了 基于 RANSAC（随机采样一致性）的标定算法，并详细介绍了将 RANSAC 算法 用于眼视线跟踪系统中的标定过程的步骤和注意事项。最后进行对 5 个测试者的 17 组实验数据进行研究分析，得出几点相关结论，其中当标定过程中出现标定点 错误的情况时，基于 RANSAC 的标定算法的性能好于基于瞬时值的标定算法， 说明了 RANSAC 算法的鲁棒性。 关键词：视线跟踪 瞳孔定位 RANSAC 标定算法 Ab Abstract 万方数据 万方数据 万方数据 万方数据 Abstract With the rise of human-computer interaction technology, research on eye gaze tracking technology is increasingly mature. The key question of eye gaze tracking system is about the users head movement problems and accuracy problems. This paper first describes the background of eye gaze tracking technology, and then introduces the characteristics of invasive and non -invasive eye -tracking technology and its development. Next, eye gaze tracking system were introduced from the calibration algorithm, accuracy issues and head movement solutions, to provide readers with some ideas. Subsequently, discusses the processing modules to optimize the parameters of the pupil extraction algorithm, based on random sampling using a 3-point RANSAC (random sample consensus ) algorithm. Finally, this paper discusses the robust calibration methods of the eye gaze tracking system First, analyzes the impact of interference points during the calibration process due to blink, gaze wander and other causes to the system performance. Then this paper proposes the RANSAC (random sample consensus) calibration algorithm, and details the steps of RANSAC algorithm for eye gaze tracking system calibration process. At last, this paper analyzes the experimental data from 17 researches carried out by five testers, and draws some conclusions ,which briefly describes is that when the marked points error hap