基于全方位视觉的自动门安全和节能检测技术  

Safety and energy saving detection technology of automatic door based on omni-directional vision sensor

作　　者：林璐璐[1] 姜荣剑 徐海涛[1] 汤一平[1] 

机构地区：[1]浙江工业大学信息工程学院,杭州310023

出　　处：《计算机应用》2014年第6期1825-1829,共5页journal of Computer Applications

基　　金：国家自然科学基金资助项目(61070134)

摘　　要：针对自动门运行过程中的效率及安全问题,提出了一种基于全方位视觉传感器(ODVS)的自动门安全和节能智能检测技术。首先,采用ODVS采集自动门周边的360°全景图,并对全景图像按检测要求进行预处理;接着,采用一种运动历史/能量图像(MHoEI)算法来检测和跟踪运动目标前景对象;然后,根据前景对象的运动方向和空间位置等信息来分析行人的行为;最后,根据行人的行为和状态来控制自动门的开启和关闭,以达到自动门安全性、节能性和舒适性的控制目标。实验结果表明,该检测方法能有效地识别自动门周围的行人行为,有助于避免各种自动门安全隐患发生,同时能高精度统计出进出自动门的人流量。Concerning the efficiency and security issues of automatic door, an safety and energy saving detection technology of automatic door based on Omni-Directional Vision Sensor （OI）VS） was proposed. Firstly, on-site 360° panorama image around automatic door was collected timely by ODVS and preprocessed according to the detection requirements. Secondly, moving target was detected and tracked through Motion History or Energy Images （MHoEI） algorithm. Then, the behavior of pedestrian was analyzed according to the direction of motion and spatial position of the foreground object. Finally, in order to make automatic door secure, energy-saving and comfortable, the automatic door was controlled to open or close according to the behavior and state of pedestrian. At the same time, the number of people passing through the automatic door can also be worked out accurately. This technology can be directly used in intelligent monitoring and business survey. The experimental results indicate that the detection technology of automatic door can recognize the behavior of pedestrian around the automatic door, avoid the security risks of automatic door and improve the accuracy of people counting.

关 键 词：全方位视觉传感器 自动门安全控制 运动历史 能量图像算法 行为检测 人流量统计 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]