基于变量选择的堆肥胡敏酸含量近红外光谱分析  

Near Infrared Spectral Modeling Analysis Based on Variable Selection of Compost Humic Acid Content

作　　者：李自刚[1,2] 岳晓禹[1] 李长滨[1] 张春辉[2,3] 魏庆葆[1] 刘世亮[4] 

机构地区：[1]河南牧业经济学院包装与印刷工程学院,郑州450011 [2]河南牧业经济学院河南省生猪疾病防控工程技术研究中心,郑州450011 [3]澳大利亚动物卫生实验室 [4]河南农业大学资源环境学院,郑州450002

出　　处：《农业机械学报》2017年第2期300-304,共5页Transactions of the Chinese Society for Agricultural Machinery

基　　金：国家自然科学基金项目(U1404332);河南省重点科技发展计划项目(112102110034)

摘　　要：为探讨基于选择的近红外光谱变量定量判别堆肥腐熟度的可能性,为堆肥发酵终点判定及开发相应的近红外漫反射光谱控制设备提供理论基础,采集了100份堆肥样本,用湿化学方法分析了堆肥样本中的胡敏酸含量,采集范围为4 000~9 000 cm^(-1)波段的光谱,进而利用该近红外光谱进行预处理方法的优化研究及堆肥胡敏酸感变量的优选研究,以偏最小二乘的方法建立了基于近红外光谱分析的堆肥胡敏酸定量模型。结果表明:小波变换充分提取了近红外光谱的信息;利用C_4小波系数对58个堆肥样本进行建模,对42个预测集样本进行预测,预测集均方根误差和R2分别为0.113 8%和0.926 6,优于原始光谱直接建模的0.167 2%和0.834 8。对近红外光谱数据进行小波变换,利用处理后的小波系数,采用偏最小二乘法预测了堆肥中胡敏酸的含量,建立了小波变换与近红外光谱技术结合用于测定堆肥样品中胡敏酸测定的模型。基于小波系数的模型优于传统的近红外光谱全谱模型,对于堆肥胡敏酸的测定取得了较为准确的预测结果。Composting treatment is a key method of processing organic solid waste, especially for agricultural organic solid waste. Aiming to study the feasibility of several selected variables from near infrared spectroscopy to quantify humic acid in compost,determine composting fermentation and develop corresponding control equipment to provide theoretical basis by near-infrared diffuse reflection spectrum.Totally 100 composting samples were collected,including 58 samples for calibration and 42 samples for validation. On the one hand,the humic acid of these samples were analyzed by using the International Humic Substances Society standard of humic acid method,on the other hand,those were scanned to obtain near infrared spectra with the wavelength range of 4 000 ~ 9 000 cm- 1. Both of spectroscopic pre-treatment method and sensitive variables were optimized,and then the model was built by partial least squares regression method. The results indicated that humic acid in compost can be determined by near-infrared（ NIR） spectral technique,because they were combined with organic groups with NIR absorption. A method for the determination of humic acid in compost samples was established based on the combination of discrete wavelet transform（ DWT） and NIR technique. In the proposed method,theraw NIR data and their wavelet coefficients were used for modeling and prediction of the contents of humic acid in compost by partial least square method（ PLS）. The model based on wavelet coefficients was better than that based on the full NIR spectral range. With the improved method,accurate prediction can be achieved.

关 键 词：堆肥 品质控制 腐熟度评价 胡敏酸 近红外光谱 小波变换 

分 类 号：S24[农业科学—农业电气化与自动化]