中文摘要：

针对现实复杂环境中,利用单个或多个热释电传感器对人体目标识别时所表现出的不稳定性和抗干扰性较差问题,提出了一种基于人体热释电特征的多策略融合识别方法。该方法以不确定性理论为基础,首先利用不同空域的传感器,获取人体不同部位的置信度,其次结合不同的融合策略实现对人体目标的识别。融合策略分为两种,第一种策略首先使用同空域融合,然后使用同时域融合,这种策略充分利用了分布式传感网络的特性,融合了最优置信度以实现对人体目标的高精度识别,第二种策略则首先使用同时域融合,然后使用同空域融合。该策略能实现对人体目标的实时识别,但不能保证较高的识别精度。实验结果表明多策略融合既能克服单传感器或多传感器在对人体目标进行识别时,所表现出的不稳定性和不可靠性,又能增强系统识别时的鲁棒性和有效性。与其他融合方法相比突出体现了本文方法的优势。

英文摘要：

Aiming at the instability and weak anti-interference characteristic in human target identification with single or multiple pyroelectric infrared（ PIR） sensors in real complex environment,this paper presents a multi-strategy fusion identification method based on human pyroelectric infrared feature. Taking uncertainty theory as the basis,this method first makes use of different spatial sensors to obtain the confidence levels in different parts of the human body,and then combines with different fusion strategies to achieve the recognition of different human targets. The fusion strategy is divided into two kinds. The first strategy uses the fusion method in the same spatial domain at first,then in the same time domain. This strategy makes full use of the characteristics of the distributed sensor network,and fuses the optimal confidence levels to realize high accuracy human target identification. The second strategy uses the fusion method in the same time domain at first,and then in the same spatial domain. This strategy can realize the real-time human target identification,but cannot guarantee high recognition accuracy. The experiment results show that the multi-strategy combination method can overcome the instability and unreliability in human target identification caused by single sensor or multiple sensors; meanwhile,the method can also enhance the robustness and effectiveness of the recognition system. Compared with other fusion methods this method shows its superiority. At the same time,comparing with other fusion method,to highlight the superiority of the proposed method