中文摘要：

为实现对玉米颗粒的自动分类,利用碰撞声信号装置采集玉米完好粒、虫蛀粒和霉变粒的840个声信号.分别从时域和频域对碰撞声信号进行分析和处理,提取信号特征.采用主成分分析方法对特征数据降维,利用BP神经网络进行分类.实验结果表明：该方法对完好粒、虫蛀粒和霉变粒3种玉米颗粒分类的正确率均达到90%以上.表明利用碰撞声信号识别玉米完好粒、虫蛀粒和霉变粒的效果良好,具有较强的实际应用价值,为检测玉米颗粒品质提供了一种新的途径.

英文摘要：

In order to realize the automatic classification of corn kernels,this approach collected 840 impact acoustic signals of undamaged kernels,insect damaged kernels and moldy kernels by apparatus of collecting impact acoustic signal,analyzed these signals from the time and frequency domain,extracted the signal features,used the principal component analysis method to reduce the dimensions of the feature data.Finally,BP neural network is used to classify the corn kernels.The classification accuracy of undamaged kernels,insect damaged kernels and moldy kernels were above 90%.The experimental results show that using impact acoustic signal,one can gain a good result in identifying undamaged kernels,insect damaged kernels and moldy kernels.So the approach has a more comprehensive value in practical application and provides a new method for corn kernels quality detection.