基于视觉的非接触式手势交互技术研究

发布时间：2019-05-21 18:53

【摘要】：随着计算机技术的进步,人机交互正在朝着人性化与简单化的方向不断发展,非接触式手势交互逐渐成为人机交互的一个重要方式。与鼠标、键盘、触摸屏等接触式的操作方式相比,非接触式的手势交互更加符合人的自然行为。近年来,随着计算机视觉和图像处理技术的发展,基于机器视觉的手势与计算机交互的方式正在成为新型人机交互方式的主流。然而,现有的手势交互体验并不理想,如何克服环境因素的干扰,准确获取手势信息并进行有效识别,如何设计一种友好、高效的手势交互方式来实现人与计算机之间的直接对话,是本文研究要解决的主要问题。本文主要的研究工作包括:(1)通过Kinect传感器获取人体手臂骨骼关节信息,提取有效手臂手关节点在三维空间的运动轨迹,并将三维运动轨迹转化为一维的手势轨迹,结合手势轨迹的表观特征,提出了一种速度与角度相结合进行手势分割的方法,进而利用改进的动态时间规整(dynamic time warp,DTW)算法实现了在复杂环境和不同光照条件下对动态手势的快速、高效识别。(2)将手关节点的空间坐标位置转化为计算机屏幕上的光标位置,从而可根据手关节点的移动来确定光标的移动方向和移动距离,实现对光标的定位。将手势识别结果转换为鼠标操作指令发送给计算机,实现手势与鼠标的映射。(3)基于MFC开发了一款手势操作五子棋游戏软件,可利用设计的手势来操作游戏,实现了一种非接触式的手势交互应用。非接触式手势交互技术解放了用户肢体及其在空间上的限制,是智能人机交互研究领域的重要组成部分。将智能科技与传统交互模式结合,开发一种新的交互技术适应了当前科技发展的趋势。本课题得到国家自然科学基金项目“No.61573029”和北京市自然科学基金项目“No.4162012”的资助,对于促进手势交互技术的广泛应用具有重要实际意义。
[Abstract]:With the progress of computer technology, human-computer interaction is developing towards humanization and simplification, and contactless gesture interaction has gradually become an important way of human-computer interaction. Compared with mouse, keyboard, touch screen and other contact operation methods, contactless gesture interaction is more in line with human natural behavior. In recent years, with the development of computer vision and image processing technology, gesture and computer interaction based on machine vision is becoming the mainstream of new human-computer interaction. However, the existing gesture interaction experience is not ideal, how to overcome the interference of environmental factors, accurately obtain gesture information and effectively identify, how to design a friendly, Efficient gesture interaction to realize the direct dialogue between people and computers is the main problem to be solved in this paper. The main research work of this paper is as follows: (1) the bone and joint information of human arm is obtained by Kinect sensor, the movement trajectory of effective arm hand node in three dimensional space is extracted, and the three dimensional motion trajectory is transformed into one dimensional gesture trajectory. Based on the apparent characteristics of gesture trajectory, a method of gesture segmentation based on speed and angle is proposed, and then the improved dynamic time warping (dynamic time warp, is used. DTW) algorithm realizes the fast and efficient recognition of dynamic gestures in complex environment and different lighting conditions. (2) the spatial coordinate position of the hand node is transformed into the cursor position on the computer screen. Therefore, the moving direction and distance of the cursor can be determined according to the movement of the hand node, and the positioning of the cursor can be realized. The result of gesture recognition is converted into mouse operation instruction and sent to computer to realize the mapping between gesture and mouse. (3) based on MFC, a kind of gesture operation Gobang game software is developed, which can use the designed gesture to operate the game. A contactless gesture interaction application is implemented. Contactless gesture interaction technology liberates user limbs and their spatial limitations, and is an important part of intelligent human-computer interaction research field. Combining intelligent science and technology with traditional interaction mode, a new interactive technology is developed to adapt to the current trend of scientific and technological development. This project is supported by the National Natural Science Foundation of China (No.61573029) and the Beijing Natural Science Foundation of Beijing (No.4162012). It is of great practical significance to promote the wide application of gesture interaction technology.
【学位授予单位】：北京工业大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP391.41

【相似文献】