Abstract: In view of the high cost and cumbersome pretreatment process of the electronic nose, the integrated fast electronic nose of MOS array, pretreatment and selfpriming injection is designed, including the preprocessing module of selfpriming headspace injection, MOS sensor array and its debugging module, as well as the data acquisition and analysis module. Based on the prototype of the fast electronic nose system, the stability of sample pretreatment, sample injection and sensor array is investigated. It is found that the relative standard deviation of manual mode, accurate selfpriming mode and simple selfpriming mode increases in turn, but the appropriate operation mode can be selected according to the characteristics and accuracy requirements of different samples. Secondly, the experiments of distinguishing the odor characteristics of three kinds of food were carried out. The results showed that the odor characteristics of fresh Meretrix of different brands of fruit juice, adulterated Baijiu and different drying time could be effectively distinguished, and the corresponding key distinguishing sensors were screened out for the above three kinds of food. Finally, the core pattern recognition algorithm of portable fast electronic nose system was designed, and an example was verified based on MATLAB. The classification accuracy was 100%.

Key words: MOS array, pretreatment, selfpriming sampler, electronic nose, odor characteristic differentiation, pattern recognition

摘要： 针对电子鼻成本高昂和烦琐前处理工序问题，设计MOS阵列、前处理以及自吸进样一体化快速电子鼻，快速电子鼻包括自吸式顶空进样前处理模块、MOS传感器阵列及其调试模块、数据采集及分析模块。以快速电子鼻系统样机为平台，首先进行样品前处理、进样以及传感器阵列的稳定性考察，发现手动模式、精确自吸模式以及简易自吸模式的相对标准偏差依次增大，但根据不同样品特点和精度要求可以选择合适的操作模式；其次，进行三类食品气味特征的区分试验，结果表明，不同品牌果汁、掺假白酒以及不同干露时间鲜活文蛤的气味特征均可以进行有效区分，并针对以上三类食品分别筛选出对应的关键区分传感器；最后开展便携式快速电子鼻系统的核心模式识别算法设计，并基于MATLAB进行实例验证，分类准确率为100%。

关键词: MOS阵列, 前处理, 自吸进样, 电子鼻, 气味特征区分, 模式识别