手持设备上基于增强现实的虚实交互技术的研究与应用

【作者】 李可歆；

【导师】 杨厚俊；

【作者基本信息】 青岛大学 ， 计算机软件与理论， 2010， 硕士

【摘要】 增强现实技术(Augmented Reality,简称AR)是在虚拟现实技术(Virtual Reality,简称VR)的基础上发展起来的一个新的研究领域,成为人机界面技术发展的一个重要方向。增强现实技术将计算机生成的虚拟对象及其相关信息准确实时的叠加到真实环境中,可以通过显示设备看到一个真实场景中无缝融入虚拟对象的虚实结合空间,为人们带来一个视觉效果更真实、场景信息更丰富的新环境。随着移动设备和技术的发展,在这类通用性高、普及率大、便携性好、能耗低的手持设备设计开发增强现实系统对增强现实技术的推进和发展具有重要的理论意义和现实意义。人机交互技术的发展将技术核心由计算机转向用户,而人手动作和手势是人类最自然、直观、易于掌握的人际交流方式,并含有丰富的交互信息,用户可以直接将手作为计算机的输入设备,通过手势发布命令,不需要交互媒介实现人机通讯,因此基于手势识别的交互技术已逐渐成为人机交互领域的研究热点。因此,研究便携式手持设备上实现增强现实手势交互技术具有很大实用价值,论文的主要内容有：1、研究了自适应三帧差分运动目标检测算法。运动目标检测是实现虚实交互的前提条件。论文研究了目前主流的视频检测算法,分析了这些算法的优势和不足,提出了一种通过运动目标速度自适应调整帧间隔的三帧差分检测方法,实验表明,与目前主流的视频检测算法相比,本文方法在运动物体方向不改变的条件下,能够快速准确的定位运动目标区域,有效降低手持设备处理延迟性,适用于计算性能较低的手持设备。2、手持设备上增强现实的虚实交互系统需要利用手持设备内置摄像头实时采集视频图像,由于使用者生理上无意识的手部抖动或移动手持设备,都可能导致视频帧背景发生变化。本文研究了Surendra背景更新算法,实现能够自适应更新背景,并与自适应三帧差分和背景差分法相结合,解决了因手持设备抖动而引起的检测误差问题,提高检测精度和效果。3、提出了基于区域相割的指尖检测算法,将交互点的范围缩小到一个特定的区域,能够确定人手指尖与虚拟物体交互的有效区域,使人手对虚拟物体的动作更加精确和协调。更多还原

【Abstract】 Augmented Reality (Augmented Reality, referred to as AR) is a new field of research developed on the basis of virtual reality technology (Virtual Reality, referred to as VR). It also is an important development direction of human-machine interface technology. Augmented reality technology accurately adds the computer-generated virtual objects and their related information to the real environment, you can see a reality-virtual space by display device which virtual objects seamlessly integrated into a real scene. It can brings a more real and new environment. As the development of mobile devices and technology, design augmented reality system on such handheld device of high universality, high diffusion rate, portable, low energy consumption will promote the development of augmented reality technology and bring important theoretical significance and practical significance.Human-computer interaction technology’s technical core shifted from computer to users. The gestures of human is interpersonal communication methods which is the most natural, intuitive and easy to master and contains much more interactive information. Users can directly use hand as a computer input device, issued an order through the gestures, does not require the media to achieve human-computer interaction. So, the interaction based on gesture recognition technology has gradually become a research highlight in human-computer interaction research field.Therefore, the study of interactive technology of AR based on handheld devices will bring great practical value. The main work in this paper is as followed:Firstly, a algorithm of adaptive three frame difference moving target detection is proposed. Moving target detection is a prerequisite for interaction. By analyzing the strengths and weaknesses of mainstream video detection methods, a new algorithm based on moving target rate which could adaptive adjust frame interval is proposed. The experiments show that this algorithm is faster than other existed methods used aimed at moving target detection. Therefore it is feasible for end-users’devices with low computational performance effectively and reduce the processing delays of handheld devices.The reality interactive system based on handheld devices needs integrated cameras capture real-time video image. As the user unconscious physiological hand shake or move mobile handheld devices are likely to result the background changes. Surendra’s algorithm which could achieve adaptive update the background, combined with adaptive three frame difference and background difference solved detection error problems be caused by jitter on hand-held device. It can improve the detection accuracy and effectiveness.更多还原