摘要
为了提高主动视觉平台的适用性、减小其复杂度,提出了一种新的主动视觉系统结构。利用摄像机作为传感器代替传统的跟踪器计算头盔的位姿。对头盔识别采用了合作目标的方式,提高了识别率和位姿计算速度。采用了一种合作目标亚像素精度定位的方法以提高合作目标三维空间的定位精度。利用PID算法对主动视觉平台进行姿态的控制。最后,实验结果证明了该方法的可行性。
To improve the applicability and reduce the complexity of active vision system, a new configuration is proposed. The orientation of helmet is got by a camera instead of the traditional tracker. The cooperative object method is adopted to recognize objects to improve the recognition rate and the speed of the orientation calculation. To get the higher locating accuracy of the cooperative object in 3D Cartesian space, a locating method with sub-pixel accuracy is implemented. The moving of the active vision platform is controlled by PID algorithm. Finally, experimental results are provided to verify the efficiency of the proposed method.
出处
《光学仪器》
2009年第4期74-78,共5页
Optical Instruments
基金
"863"国家高科技资助项目(2006AA04Z228)
国家自然科学基金资助项目(60675045)
关键词
主动视觉
合作目标
识别
亚像素
active vision
cooperative object
recognition
sub-pixel
