动态手势轨迹跟踪人机交互系统毕业论文.docxVIP

PAGE 5 本科毕业设计(论文) 题 目: 动态手势轨迹跟踪人机交互系统 姓 名: 孙超 专 业: 机械设计制造及其自动化（机械电子） 班 级: 1003014 学 号: 100301415 指导教师: 汪希伟 职称: 讲师 二○一 四 年 五 月 十五 日 动态手势轨迹跟踪人机交互系统 摘 要 人与计算机的交互活动变得越来越密不可分，近些年人机交互技术变得异常活跃，然而以手势为基础的人机交互更成为人们研究中的热点。本文基于机器视觉技术开发了一套人机交互系统，实现用户动态手势采集与识别，向计算机发布指令。通过动态手势的轨迹追踪实现人与计算机的实时交互，更加自然并且效率更高。 本文设计并采集了多组动态手势。基于HALCON图像处理软件，通过颜色及形状特征进行手部图像分割。记录手部某一基点坐标在连续图像中的变化，得到动态手势的运动轨迹。提取运动轨迹的（周）长宽比、（周）长高比、宽高比、边数等特征，统计不同动态手势的轨迹特征，对动态手势进行建模识别。测试结果表明对数字0、1、2、4、符号√动态手势各五组的识别率达到100%，表明该方法可以对所设定的动态手势轨迹成功识别。 在预存的手势图像识别的基础上对实时图像进行动态采集与分析处理，通过HALCON连接笔记本内置摄像头采集图像，用基于预存图像开发的处理程序处理动态图像，可以识别出每种动态手势轨迹。但由于受光照强度、环境复杂程度等因素动态变化的影响，成功率不及对预存图像处理的结果。 关键词：动态手势；轨迹追踪；人机交互 Dynamic gesture trajectory tracking human-computer interaction systems ABSTRACT The interactions between human and computers are becoming more and more inseparable. In recent years, the technology used in the interactions has become extremely active. However, the human-computer interaction based on the gesture has become a hotter spot in the study of human being. In this research, based of machine vision technology ,a human-computer interaction system is developed, which can collect and identify the users dynamic gestures, then issues instructions to the computer. Implement real-time interaction of human and computer is tracked through dynamic gesture trajectory. Its more nature and efficient. In this research, plenty of dynamic hand gestures are designed and collected. According to HALCON image processing software, we can cut up the image of hand by color and shape features. Recording to the change of the hand of a basis point coordinates in continuous image. Then get the trajectories of dynamic gestures. To extract the trajectory of the ratio of the circumference and width、the ratio of the circumference and height、the ratio of the wi