基于视觉的手势识别技术在车载主机上的应用  

On the host car application of gesture recognition technology based on vision

作　　者：陈义 马云林 

机构地区：[1]广东好帮手电子科技股份有限公司,广东佛山528100

出　　处：《电子设计工程》2016年第8期141-144,共4页Electronic Design Engineering

摘　　要：手势作为一种自然、直观、易于学习的人机交互手段,手势输入是实现自然、直接人机交互不可缺少的关键技术,目前的手势识别技术主要分为基于数据手套和基于视觉两种。基于在车载主机上实现手势识别技术来保障行车安全的目的,提出了一种基于视觉的手势识别技术在车载主机上的应用,通过红外摄像头和Kinect传感器捕获有效手势图像帧,并结合1S内连续6幀图像中手势的相对位置确定手势运动方向,大大降低了手势误判率。通过实际市场反馈表明,手势识别技术在车载主机的成功运用,不仅提高了车主的驾驶体验感,还保障了行车安全,取得了良好的效果。Gestures as a natural, intuitive and easy to learn means of human-computer interaction, signal input is to realize the natural, direct indispensable key technology of human-computer interaction, the gesture recognition technology based on vision is mainly divided into based on data glove and two. Based on gesture recognition technology on vehicle host to ensure driving safety, the purpose of this paper proposes a gesture recognition based on vision technology in the car on a host of applications, through the infrared camera and capture device sensors effectively gesture image frames, and connecting with the 1s in six consecutive frames in the gesture of relative position determine direction, greatly reduces the gesture misjudgment rate. Through the actual market feedback shows that host the successful use of gesture recognition technology in the car, the driving experience not only improve the owners, also guarantee the safety, good results have been achieved.

关 键 词：手势识别 人机交互 车载主机 Kinect传感器 行车安全 

分 类 号：TN302[电子电信—物理电子学]