基于金属多特征联合的硬币识别器设计  

DESIGNING COIN RECOGNISER BASED ON MULTIPLE METAL CHARACTERISTICS COMBINATION

作　　者：鲁彦玉 刘振兴[1] 

机构地区：[1]武汉科技大学,湖北武汉430081

出　　处：《计算机应用与软件》2015年第11期222-225,共4页Computer Applications and Software

基　　金：国家自然科学基金项目(61174107)

摘　　要：随着公共场合投币收费服务方式的流行,需要自动识别硬币设备的应用越来越多。为了提高硬币识别准确性、假币的拒识率以及增强识币设备对温度等外界因素影响的鲁棒性,基于金属的多种特征识别,设计了一种硬币识别器。方案首先通过大量的实验采集硬币的三种主要特性数据(电导率、磁导率和面积大小),然后通过实验数据分析建立硬币的联合概率空间,最后通过检测样本的特征,并计算特征参数是否处于硬币的联合概率空间内来判断样本的真伪和面值。对样本的判定算法进行优化,使之更适用于单片机处理。实验结果表明,所设计的识币器准确率较高,鲁棒性较强。With the popularisation of coin fee-based service in public places, the applications for the need of automatic coin detection de- vices are growing. To improve the coins recognition accuracy and the rejection rate of counterfeit ones, and to enhance the robustness of the coin detector against the influence from external factors such as temperature, etc. , this paper designs a coin recogniser based on the identifi- cation of multiple characteristics of metal. First, the scheme collects three main characteristics data of coins by a large number of experi- ments, which are respectively related to the electrical conductivity, permeability and area size. Then it builds a joint probability space of coins through analysing the experimental data. Finally, by the detecting the characteristics of sample and calculating the characteristics pa- rameters, it estimates the authenticities and the denominations of the sample by whether these parameters are within the joint probability space of coins or not. The paper also optimises the sample discrimination algorithm to make it more suitable for MCU to implement. Experimental result demonstrates that the designed coin recogniser has higher accuracy and robustness.

关 键 词：硬币识别 联合概率空间 准确度与鲁棒性 

分 类 号：TP212.9[自动化与计算机技术—检测技术与自动化装置]