基于Kinect的增强现实人手康复自然交互

发布时间：2018-06-15 13:10

  本文选题：增强现实 + Kinect　； 参考：《南京信息工程大学》2016年硕士论文

【摘要】：增强现实(Augmented Reality,简称AR),是在虚拟现实基础上发展起来的新技术。它是将计算机生成的虚拟物体、场景或提示信息叠加到真实场景中,从而实现对现实的增强。近年来,增强现实技术在康复训练中逐渐得到应用。将AR技术应用于手部康复,除了具有虚拟现实(Virtual Reality,VR)康复技术的优点之外,更具感知优势。身处其中的用户并未与周围的真实环境相隔离,他们能够在场景中看到自己真实的手,并以更加自然的方式与真实场景及虚拟物体进行实时交互,提高了交互的真实性与实用性。本文以国家自然科学基金项目为背景,将EMD算法引入并指等复杂手势识别,并开发了一个基于Kinect的增强实现手部康复自然交互系统。本文主要研究内容包括以下几个方面：首先本文通过利用Kinect设备,在静态识别下,针对手部康复过程中手指粘连或并拢等复杂问题提出了EMD算法准确地识别手势,并用这种算法和传统的基于模块、向量机等识别算法进行比较,详细分析了各种算法的优缺点,证明基于EMD手势识别算法的优越性。针对动态手势的增强现实,本文基于Openframeworks框架建立了增强现实交互环境。其中增强现实交互为二维交互和三维交互,基于二维的交互主要将虚拟物体直接根据二维坐标叠加到场景中,进行交互。而基于三维的交互先获得内外参数,然后得到创建的三维坐标系,最后将三维模型载入,实现互动。本文在相关医生的指导下,依照手部作业治疗的康复训练方法设计了抓握捏训练、轨迹训练、单手交互训练和双手交互训练项目,这些训练项目在难度等级上存在递进关系,最后通过实验对参数值指标进行问卷评估和分析,验证了本系统的有效性和可行性。
[Abstract]:Augmented reality is a new technology developed on the basis of virtual reality. It overlay the computer generated virtual object, scene or prompt information into the real scene, so as to realize the enhancement of reality. In recent years, augmented reality technology has been gradually applied in rehabilitation training. The application of AR technology in hand rehabilitation has the advantages of virtual reality (VR) and virtual reality (VRV). Instead of being isolated from the real environment, the users are able to see their real hands in the scene and interact in real time with the real scene and virtual objects in a more natural way. It improves the authenticity and practicability of interaction. Based on the project of National Natural Science Foundation of China, this paper introduces EMD algorithm into hand recognition and develops an enhanced hand rehabilitation natural interaction system based on Kinect. The main contents of this paper include the following aspects: firstly, by using Kinect equipment and static recognition, an EMD algorithm is proposed to recognize hand gestures accurately, aiming at the complex problems of finger adhesion or convergence during hand rehabilitation. Compared with traditional algorithms based on module and vector machine, the advantages and disadvantages of these algorithms are analyzed in detail, and the superiority of gesture recognition algorithm based on EMD is proved. Aiming at the augmented reality of dynamic gesture, this paper establishes an augmented reality interactive environment based on Openframeworks framework. The augmented reality interaction consists of two dimensional interaction and three dimensional interaction. The interaction based on two dimensions mainly superposes the virtual object into the scene directly according to the two dimensional coordinates. The internal and external parameters are obtained based on the 3D interaction, and then the 3D coordinate system is created. Finally, the 3D model is loaded to realize the interaction. Under the guidance of relevant doctors, according to the rehabilitation training method of hand work therapy, this paper designs grip training, track training, one-hand interactive training and two-hand interactive training. These training items have a progressive relationship in the level of difficulty. Finally, the validity and feasibility of the system are verified by the questionnaire evaluation and analysis of the parameter index.
【学位授予单位】：南京信息工程大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：R49;TP391.9
，

本文编号：2022115