基于图像处理的半虚拟键盘的研究

发布时间：2018-03-13 02:25

本文选题：半虚拟键盘　切入点：图像处理　出处：《华南理工大学》2013年硕士论文　论文类型：学位论文

【摘要】：随着现代计算机技术的发展，诸如笔记本、智能手机、平板电脑等便携式设备日益趋向功能集中化、小型化，而传统键盘存在体积大、不方便携带的缺陷，手机键盘又限于按键尺寸过小不方便操作，这些阻碍了人机交互的舒适性与便携性。由此，一种基于图像处理技术，具有硬件少、体积小、方便携带等优点的虚拟键盘成为解决上述问题的新型人机交互工具。基于图像处理技术的虚拟键盘，不需要键盘实体，通过摄像头采集键盘区域图像，经图像处理识别出按下的按键和用户击键的动作，用软件产生按键命令，实现键盘的功能。在本文中，将按键通过投影产生的虚拟键盘称为纯虚拟键盘，将按键通过打印或丝印在其他物体上的虚拟键盘称为半虚拟键盘。从虚拟键盘的实现上讲，其主要解决三个问题：键盘位置的定位，按键的识别和按键按下与释放的模态识别。通过键盘4个顶点定位键盘区域的方法不够准确，影响按键的识别精度，为此，本文提出了离线建模的方法，能够定位到键盘的每一行。采用手势识别、模型匹配或阴影分析等方法实现手指定位及击键动作的识别，其算法受外界光线影响较大，效率不高，，很难在实际中应用。为此，本文采用红外激光作为辅助光源实现按键的定位及模态的识别。本文首先介绍了虚拟键盘的国内外研究状况，在此基础上阐述了本文的设计方法，介绍了半虚拟键盘系统。通过离线建模获取键盘行列位置信息，以光斑定位指尖位置，再利用键盘行列信息识别目标按键，通过比较相邻两帧图像的按键信息识别按键的模态，最终实现了键盘功能。最后对离线建模及按键效果做了实验与分析，实验表明本文设计的半虚拟键盘能够快速、高准确度识别按键及其模态，受外界光线影响小，适于在实际中使用。
[Abstract]:With the development of modern computer technology, portable devices such as notebook, smart phone, tablet computer and so on tend to be centralized and miniaturized. The keyboard of mobile phone is limited to too small keystroke size to operate easily, which hinders the comfort and portability of human-computer interaction. Virtual keyboard, which is easy to carry and so on, has become a new human-computer interaction tool to solve the above problems. The virtual keyboard based on image processing technology does not need keyboard entity. The image of keyboard area is collected by camera, and the keystroke action is recognized by image processing, and the key command is generated by software. In this paper, the virtual keyboard generated by the key projection is called the pure virtual keyboard, and the virtual keyboard that the keys are printed or printed on other objects is called the semi-virtual keyboard. From the realization of virtual keyboard, it mainly solves three problems: the location of keyboard, the identification of keystroke and the modal identification of press and release of keystroke, the method of locating keyboard area through four vertices of keyboard is not accurate enough. This paper presents an off-line modeling method, which can locate every line of the keyboard. Hand gesture recognition, model matching or shadow analysis are used to realize finger positioning and keystroke recognition. The algorithm is greatly influenced by the external light, and its efficiency is not high, so it is difficult to be applied in practice. Therefore, the infrared laser is used as the auxiliary light source to locate the keys and identify the modes. This paper first introduces the research status of virtual keyboard at home and abroad, and then expounds the design method of this paper, introduces the semi-virtual keyboard system, obtains the position information of keyboard row and column by off-line modeling, and uses spot to locate finger tip position. Finally, the keyboard function is realized by comparing the modes of the keystroke information of the adjacent two frames. Finally, the off-line modeling and the effect of the keystroke are tested and analyzed. The experiments show that the semi-virtual keyboard designed in this paper can identify the keys and their modes quickly and accurately, and is less affected by the external light, so it is suitable for practical use.
【学位授予单位】：华南理工大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP391.41;TP334.23

【参考文献】