【作者】 李新； 田斌； 辛阳； 陈林顺；

【Author】 Li Xin~1 Tian Bin~1 Xin Yang~1 Chen Linshun~2 (1.Key Laboratory of network and information attack and defence technology of Ministry of Education, Beijing University of Posts and Telecommunications,Beijing 100876; 2.Beijing Safe-Code Technology Co.,Ltd.,Beijing 100876)

【机构】 北京邮电大学网络与信息攻防技术教育部重点实验室； 北京安码科技有限公司；

【摘要】 近年来,异常事件的检测已经成为传感器网络领域中的研究热点。我们针对传感器网络中音频异常事件检测这一问题,提出了一种基于贝叶斯决策的音频异常事件检测方法。我们通过建立环境池来存储正常事件并对正常的情况进行建模,并根据时间、地点等因素对环境池进行动态的调整。通过环境池,我们可以知道采集到的音频信号与正常事件的相似程度,在此基础上使用贝叶斯决策可以对异常事件进行判断。我们将该方法部署在音频传感器网络中进行验证,实验结果证明,本文提出的方法能够有效的检测出传感器网络中的音频异常事件。更多还原

【Abstract】 In recent years,abnormal event detection has been paid much attention in the field of WSNs.In order to detection the abnormal events in the audio sensor networks,in this paper,we present an audio -based abnormal events detection method for wireless sensor networks(WSNs).In the proposed method,we use a normal event pool,which is called Environment Pool,to store the models of the normal audio events.Specially,we update the Environment Pool according to some environment parameters such as time,location, and so on.Through comparing the current audio signal with the Environment Pool,we can get the similarity between the collected audio signals and normal events.Based on the similarity,we introduce the Bayesian decision to judge the abnormal events.We deploy the proposed method in the WSN to evaluate its performance,and the experimental results show that our method can achieve satisfying results.更多还原