中文摘要：

提出了一种基于主元分析（PCA）和核独立主元分析（KICA）双空间特征提取方法，PCA—KICA方法．运用碰撞声装置采集玉米完好粒声信号、虫蛀粒声信号、霉变粒声信号，首先将信号在PCA空间进行特征提取，然后将提取的特征送入到KICA空间提取信号的特征向量，最终送入到粒子群优化的支持向量机分类器中进行分类．实验结果证明，单空间特征提取算法对于3类信号的分类效果不理想，但是采用双空间中特征子空间的互补性可以克服单空间的限制．PCA-KICA双空间特征提取方法的识别率最高，完好粒、虫蛀粒、霉变粒的识别率分别达到95．00％、96．40％、97．80％．

英文摘要：

In this paper, a dual space feature extraction method was put forward based on principal component analysis （PCA） and kernel independent principal component analysis （KICA）, the PCA-KICA method.Irapact acoustic signal device was used to collect signals of undamaged corn kernels, insect-damaged kernels and mildew-damaged kernels.First, the samples were performed to the PCA space for feature extraction and then performed to the KPCA space for feature extraction. Subsequently, these features were input to support vector machine ,which was optimized by particle swarm optimization.Experimental results showed, single subspace could not get a better classification accuracy rate, but applying dual space feature extraction method could overcome the limits of single subspace. PCA- KICA method emerged the highest recognition rates, which were 95.00%, 96.40% ,97.80% for undamaged kernels,insect-damaged kernels and mildew-damaged kernels respectively.