摘要
利用Intelnet跨平台计算机视觉库OpenCV和微软Visual Studio 2010构建体感遥控器的软件开发平台。以OpenCV为基础,结合自定义算法和函数,开发利用人体手势识别来控制不同设备的体感遥控器。该设计采用基于codebook背景模型的运动检测方法对手势定位,以改进的CamShift算法和Kalman滤波器进行手势跟踪,进而对手势分割和识别。研究利用单目摄像头下对单手手势识别,以红外线和无线通信方式对目标区域和设备进行非接触控制。给出软件设计的基本思想,用OpenCV实现算法的具体过程和关键代码。该方案目标跟踪快、检测效率高、实时性好,为下一步在嵌入式系统中实施此方案打下一定基础。
Through Intelnet＇s cross-platform computer vision library OpenCV and Microsoft Visual Studio 2010, the motion sensing remote controller development platform is constructed. With custom algorithms and functions based on OpenCV, a motion sensing remote controller is developed by human gestures recog- nition techniques. This design adopts motion detection with codebook background model and object track- ing with improved CamShift algorithm and Kalman filter being used, to perform segmentation and recogni- tion. Single palm gesture recognition with a single webcam camera is realized to control the devices and target area through infrared and wireless communication without physical contact. The basic ideas of the design, the specific procedures for implementing algorithms of OpenCV and the key codes are given. This design has merit in fast tracking, high efficient detecting and real-time synchronizing, which lays the foun-dations for the advanced implementation in embedded systems.
出处
《黑龙江工程学院学报》
CAS
2012年第1期74-77,共4页
Journal of Heilongjiang Institute of Technology
基金
国家自然科学基金资助项目(61075039)
关键词
OPENCV
体感
遥控器
手势跟踪
手势识别
计算机视觉
OpenCV
motion sensing
remote controller
hand gesture tracking
hand gesture recognition computer vision
