中文摘要：

为建立一种无损检测桃果实瘀伤的方法,对桃果实施加外力使其产生瘀伤,在24℃、相对湿度85%条件下贮藏24 h后逐个提取电子鼻响应信号,并对其进行主成分分析。结果显示：不同瘀伤等级果实的分离度为100%。电子鼻传感器阵列所含的10个传感器响应信号与瘀伤等级的相关性分析表明,大部分传感器（传感器W5S、W6S除外）响应信号与瘀伤等级均有显著相关性。利用向后消去法进行多元线性回归分析,结合取整函数建立了桃果实瘀伤等级的预测模型,验证试验表明,该模型具有较好适用性,总体准确率达到95%。

英文摘要：

A new method to detect bruise in peaches by E-nose was proposed.Bruise damages of peaches were made by external forces and the damaged peaches were held at 24 ℃,RH 85% for 24 h.They were then tested by the E-nose one by one.Principal component analysis of sensors′ response showed that the E-nose could be used to differentiate peach fruits subjected to different external forces.The correlation analysis of 10 sensors′ response and bruise levels showed that most of the sensors′ response（sensor W5S,W6S excluded）had significant correlation with the bruise levels.A model to forecast peach bruise was established by backward elimination multiple linear regression analysis and the rounding function.The confirmation experiment indicated that the model had good adaptability on peaches with different bruise levels,and the comprehensive accuracy rate was 95%.