基于支持向量机及粒子群算法的腊肉品质等级检测

Predication of Chinese Bacon Quality Grades Based on Support Vector Machine and Particle Swarm Optimization Algorithm

针对近年来备受关注的腊肉酸价和过氧化值超标、褪色、出油、发黏等品质问题,提出一种快速、准确、实用的检测技术。采用支持向量机(support vector machine,SVM)将近红外光谱(near infrared spectroscopy,NIR)检测到的酸价、过氧化值、挥发性盐基氮和显微图像处理得到的微生物菌落总数进行多数据融合,建立腊肉品质等级检测模型,并利用粒子群优化(particle swarm optimization,PSO)算法进行模型优化。结果表明:支持向量机的分类方法取得了与生化方法相同的腊肉分级预测结果,且采用粒子群优化后的分类模型准确率由97.5%提升到100%。证明粒子群优化支持向量机模型能够迅速对腊肉等级进行准确检测。

In recent years,quality problems of Ch inese bacon such as acid values and peroxide values exceeding the national standard,color fading,oil exudation and sticky feeling to the touch have received growing attention.With that in mind,a fast,accurate and practical detection method to evaluate Chinese bacon quality is present ed in this paper.We established a predictive model for bacon quality detection by using the support vector machine（SVM）approach based on the near-infrared spectral data（acid value,peroxide value,volatile base nitrogen）and microscopic image data（the total number of microbial colonies）.Moreover,the model was optimized by using particle swarm optimization（PSO）algorithm.It was found that the prediction results of the SVM model and the biochemical method were consisted for bacon quality classification.Besides,the predictive accuracy of the classification mode was increased from 97.5% to 100% after optimization.The SVM model optimized by PSO proved to be able to quickly and accurately detect Chinese bacon quality.