基于热释电红外传感器的车辆和人员分类

设计了一种应用于野外环境中的热释电红外系统,并创新性地提出用单热释电红外传感器节点进行车辆和人员的分类.分类算法利用信号的幅度、信号的持续时间、信号一阶差分值的绝对值的最大值作为特征向量,利用支持向量机（SVM）进行目标分类.在10,20 m处分类准确率分别可以达到93.5％,94.5％,并且针对10,20m的综合分类准确率可以达到92％.该系统扩展了热释电传感器的应用范围和应用场景,所用分类方法对于其他同类传感器系统具有一定的借鉴意义.

Classification of vehicle and personnel based on pyroelectric infrared sensor

Design a pyroelectric IR （ PIR ） system applied in outdoor environment, and innovatively propose a method using single PIR node to classify vehicles and personnel. The classification algorithm uses signal amplitude, duration of signal, the maximum value of absolute value of the first order difference as feature vector, and use support vector machine（ SVM）for target classification. Classifying aeeuracy is 93.5 % and 94.5 % when PIR detectors are deployed 10 m and 20 m away from the area of interest, and the compositive classification accuracy is 92 %. The system extends the application range and scenarios of pyroeleetric sensor and the proposed classification method has a certain reference meaning for other similar sensor systems.