论文部分内容阅读
近年来 ,对将传感器阵列技术和模式识别技术用于识别食品挥发气味的研究方兴未艾。模式识别技术的实施主要依赖于对传感器阵列输出信号的参数表达 ,迄今为止 ,不论是单个传感器还是传感器阵列均没有通用的参数选择的方法。该文从 8个氧化锡气敏传感器与食醋气味反应中提出初始特征值 ,采用一种基于公式表达树的遗传基因块代码的编码算法的组织特征参数法 (OF P)对所提取的特征参数进行融合 ,从而得到最优的组织特征参数很容易区分不同的气味。它不但解决了怎样得到最优特征参数的问题 ,而且用这种算法进行遗传运算更直观、更方便。并详细讨论了怎样进行基因编码来融合不同特征参数 ,同时对这种遗传算法怎样进行选择、交叉、变异进行了研究。将其用于气敏传感器阵列对不同食醋识别的应用实例证明 ,该方法是一种非常有效的模式识别方法
In recent years, research on sensor array technology and pattern recognition technology for identifying the volatile odor of food is in the ascendant. The implementation of pattern recognition technology mainly depends on the parameter expression of the output signal of the sensor array. So far, there is no universal parameter selection method for either a single sensor or a sensor array. In this paper, the initial eigenvalues were proposed from the reaction of eight tin oxide gas sensors with the odor reaction of vinegar. A histogram based on the genetic algorithm (OF P) The parameters are fused so that the optimal tissue characteristic parameters are easy to distinguish between different scents. It not only solves the problem of how to get the optimal characteristic parameters, but also uses this algorithm to carry on the genetic operation more intuitive and more convenient. And discussed in detail how to carry on the gene coding to fuse different characteristic parameters, at the same time how to choose, cross and mutation of this kind of genetic algorithm. The application example of gas sensor array for different vinegar identification proves that this method is a very effective pattern recognition method