摘要
利用人体自然属性来实现增强现实技术是目前的研究热点之一.本文提出了将V-系统这一全新的正交函数系应用到手势识别问题中,以此实现人机交互效果.然后再将手掌看作坐标平面,手指的边缘线看作直线特征,并混合手掌的点特征,提出了基于点线融合的相机姿态估计算法,以此实现虚实融合的增强效果.基于这两步算法,搭建了基于手势识别和手掌定位的增强现实系统,实验结果表明了算法的可行性.
It is one of the research issues to realize augmented reality technology by using human natu-ral features.By introducing V-system which is a new orthogonal function system,this paper proposes a new gesture recognition algorithm to realize the human-computer interaction.Here,the palm is looked as the coordinate plane,then the pose of the camera is estimated by fusing the point features with the line features on the palm.Furthermore,an augmented reality demo system is developed with our method,and the experimental results prove the feasibility of the algorithm.
作者
张彩霞
徐然然
ZHANG Caixia;XU Ranran(Institute of Image Processing and Pattern Recognition,North China Univ.of Tech.,100144,Beijing,China)
出处
《北方工业大学学报》
2019年第5期62-68,共7页
Journal of North China University of Technology
基金
国家自然科学基金项目“P3P问题解分布的临界曲面研究”(61503004)
关键词
增强现实
V-系统
手势识别
姿态估计
augmented reality(AR)
V-system
gesture recognition
pose estimation