基于遗传组合网络的肉用人工嗅觉系统_孙永海.pdfVIP

37 　5 吉林 大学 学 报 (工学 版 ) 　Vol. 37　No. 5 2007 9 Journal of Jilin University (Engineering and Technology Edition) 　Se t. 2007 1 2 2 孙永海 ,孙钟雷,李　宇 (1. 吉林大学 生物与农业工程学院, 长春 130022;2. 长江师范学院生命科学系, 重庆408003) 　:根据肉的气味研制了 套适合肉品新鲜度检测的肉用人工嗅觉系统。通过分析肉品 腐败机理, 合理选用了气敏传感器阵列, 优化设计了仿生气味收集装置及信号调理电路。依据 肉气味模式,采用遗传优化的组合RBF 神经网络作为模式识别方法。理化试验结果表明,肉 用人工嗅觉系统检测肉新鲜度准确率达95%, 优于其他通用人工嗅觉系统。 :食品加工技术;肉品新鲜度;人工嗅觉系统;遗传算法;组合RBF 神经网络 :TS251. 7　　:A 　　:1671-5497(2007)05-1209-05 Meat artificial olfactory system based on composite GA-RBF network 1 2 2 Sun Yong-hai ,Sun Zhong-lei , Li Yu (1. Collegeof Biological and Agriculture Engineering, Jilin University, Changchun 130022, China;2. Departmentof Lif eScience, Yangtze ormal University, Chongqing 408003, China) Abstract:An artificial olfactory system w as develo ed for monitoring meat freshness according to meat smell. By analyzing the rottenness mechanism of meat, an array of gas sensors w as selected;a device for smell collection and a circuit for signal dis osal were designed. The com osite GA-RBF netw ork w as a lied in this artificial olfactory system for attern recognition. Ex eriments w ere conducted and results show that the recognition rate of this system is 95%, w hich is higher than that of other systems. Key words:food rocessins technology;meat freshness;artificial olfactory system;genetic algorithm; com site RBF neural netw ork 　　, [1- 5] [1] 。Natale 。 8 PCA 、SOM , , 。, [2] , 、。Grigioni 32 , PCA