摘要
提出一种用于摄像机室内移动物体定位的超声波传感器阵列,通过机器人学习的可编程软硬件来实现的系统。该传感器阵列可被编程设定为11对发送/接收器,通过传感器阵列初始化时所采集的静态环境数据与被定位对象进入静态环境时所采集的数据作比较,结合不可能概率事件与采样数据的正态分布及几何学的相交定位思想,以最大似然估计、最小方差计算等作为编程机制,得到水平定位极坐标。实验表明:其精准定位水平方向角为7.5°,定位目标更新周期为200 ms,1.2 s内锁定目标。并实现灵活变倍与聚焦,满足人们对于室内摄像机的定位要求。
An ultrasonic sensor array using for indoor camera position of the moving objects is proposed, and it is realized by programmable software and hardware for robot learning. The array can be programmed to 11 pairs of transmission/receiver sensor. By the way of comparing the data collected in both static environment and unstable environment, combining with the impossible probability of the events and normal distribution of the sampled data and intersecting location ideological of the geometry. The system can compute the polar orientation based on the programming mechanisms of the maximum likelihood estimation, and the weighted average operator, as well as the calculation of minimum variance. Experiments show that the precise angle of the horizontal positioning up to 7.5°. Flexible zoom and focus is realized. It can meet the requirements of indoor camera positioning.
出处
《传感器与微系统》
CSCD
北大核心
2011年第1期79-82,共4页
Transducer and Microsystem Technologies
基金
国家自然科学基金资助项目(60772027)
