利用近红外光谱法测定玉米籽粒含油量的研究

用近红外光谱(NIRS)分析技术和偏最小二乘法(PLS)建立了玉米籽粒含油量分析数学模型,并对模型预测结果的准确性进行了评价。结果表明,定标集、检验集的预测值与化学测定值间均呈极显著正相关,相关系数分别为0.958和0.957,定标标准差和预测标准差分别为0.757和0.745。利用该技术能测定玉米籽粒含油量。     

植烟土壤中全氮和有机质含量的近红外快速分析

应用傅立叶变换近红外漫反射光谱仪(NIRS)对植烟土壤样品中全氮、有机质含量进行检测,通过应用化学计量学方法,建立近红外光谱与土壤样品中全氮、有机质含量间关系的数学模型。结果表明:所建模型稳定性好,NIRS法测定的结果与化学方法测试结果相吻合,可应用傅立叶变换近红外漫反射光谱仪来快速预测植烟土壤样品中的全氮和有机质含量。     

Determination of thermal and retrogradation properties of rice starch using near-infrared spectroscopy

Starch is a major component of rice grain and thus plays an important role in grain quality. For breeding rice with improved quality, the thermal and retrogradation properties of starch may be routinely measured. Since direct measurement is time-consuming and expensive, rapid predictive methods based on near-infrared spectroscopy (NIRS) can be applied to measure these quality parameters. In this study, calibration models for measurement of thermal and retrogradation properties were built from the spectra of grain and flour samples. The results indicated that both grain and flour spectra could give similar accuracy (r²=∼0.78) in determining the peak temperature (T_p) and conclusion temperature (T_c) of gelatinization. However, flour spectra (r²=0.80) were superior to the grain spectra (r²=0.73) in measuring onset temperature (T_o). Furthermore, the thermal properties of width at half peak height (ΔT_1/2) and enthalpy of gelatinization (ΔH_g), and retrogradation properties of enthalpy of retrogradation (ΔH_r) and retrogradation percentage (R%) could only be successfully modeled with the flour spectra. The models reported in the present study are usable for routine screening of a large number of samples in early generation of selection in breeding programs. For accurate assay of the thermal and retrogradation properties, however, direct instrumental measurement should be employed in later generations.     

NIR-spectroscopic investigation of foliage of ozone-stressed <Emphasis Type="Italic">Fagus sylvatica</Emphasis> trees

Near infrared (NIR) reflectance spectroscopy was tested as a fast method for characterizing the toxic effects of air pollution on trees. Fagus sylvatica was exposed to known, different levels of ozone during summer of 2003 in plant chambers. Leaves were taken from the plants and NIR spectra were recorded. In order to derive calibration models, reference analyses were carried out and soluble carbohydrates were determined by an enzymatic method. Minor components, like α-tocopherol, were measured after solvent extraction using gas chromatography-mass spectrometry. A partial least squares (PLS) algorithm was used to perform the chemometric analyses. Good or in some cases very good calibration statistics expressed in terms of root mean square error of cross validation and R² were obtained for glucose, fructose, sucrose, and α-tocopherol. In addition, a better than expected correlation between the NIR spectra of the leaves and the ozone concentration in the plant chambers was observed. NIR spectroscopy appears to be a very useful method to simultaneously determine many components in leaves that are related to the health of trees. It is obviously able to quantitatively describe the changing patterns of constituents in leaves of trees caused by toxic substances. Thus, NIR spectroscopy can be a very effective tool for environmental biomonitoring, especially for observation of forests.     

NIRS结合TQ软件对氯化铵掺假牛奶定量分析

[目的]应用近红外光谱法建立氯化铵掺假牛奶定量分析模型。[方法]氯化铵是提高牛奶中含氮量的典型掺假物质,样品直接使用近红外光谱仪采用漫反射和三氯乙酸预处理后使用透射模块分别扫描并建立定量分析模型,并对模型进行验证。[结果]建立了漫反射氯化铵含量定量分析模型和透射氯化铵含量定量分析模型,后者模型更加准确可靠,均方根校正标准差(RMSEC)、相关系数(R2)、均方根预测标准差(RMSEP)分别为0.032 4、0.998 4、0.049 8,回收率为107.607 4%。[结论]三氯乙酸预处理后的透射模型更加精确,可以用于牛奶中氯化铵掺假检测,为进一步研究牛奶中其他物质掺假检测提供借鉴。     

基于PDS算法的不同光照下模型传递研究

针对在室外光照对样品使用近红外光谱检测带来误差的问题,提出基于模型传递来减少检测误差的方法。以圆黄梨为样品,分析样品在室内、室外阴影下的近红外光谱,建立室内光谱的PLS模型。采用分段直接校正(Piecewise Direct Standardization,PDS)算法,减小室内外光谱差距,使得室内PLS模型能预测室外光谱。结果表明,在室内建立的模型能预测经PDS算法传递后的室外光谱,预测决定系数(R2)和标准差(Root Mean Square Error of Prediction,RMSEP)分别为0.64和0.565 27,能有效地解决室外光照对光谱检测影响的问题。     

Harvest timing and N application rate effects on switchgrass yield,nutrient cycling,and partitioning

Harvest time (HT) and nutrient management may have an effect on switchgrass (Panicum virgatum L.) yield and nutrient cycling. Experimental objectives were to quantify HT and nitrogen (N) rate effects on dry matter yield (DMY), nutrient concentration, and N use. “Alamo” switchgrass was grown under four N rates (0, 40, 80, and 120 kg ha^?1) with six monthly HT (May to October) in a randomized complete block design experiment with a split-plot arrangement. Yield increased cubically and quadratically with HT in Years 1 and 2. Aftermath yield decreased linearly with HT in both years. N rate increased yield in Year 2 only with no yield benefit at rates >40 kg N ha^?1. Nutrient concentration decreased from May to October, while its removal was determined by DMY. N use and recovery were greatest at 40 kg N ha^?1 and declined with additional N inputs. Results indicate that harvesting post-frost-kill produced greater DMY with less nutrient concentration and removal.     

鱼粉中氨基酸近红外光谱定量分析

收集了145个鱼粉样本,应用偏最小二乘（PLS）方法,建立了鱼粉中17种氨基酸和总氨基酸NIRS定标模型。天冬氨酸、蛋氨酸、赖氨酸、苏氨酸、谷氨酸、甘氨酸、丙氨酸、缬氨酸、异亮氨酸、亮氨酸、苯丙氨酸、精氨酸、脯氨酸和总氨基酸的化学分析值与NIRS预测值的决定系数R^2都达到0．87以上,相对标准差均小于10％,相对分析误差均大于3;酪氨酸的决定系数R^2为0．8678,相对标准差为8．65％,相对分析误差为2．77;组氨酸、丝氨酸和半胱氨酸的决定系数R。分别为0．9005、0．7436和0．3541,相对标准差分别为14．19％、17．85％和33．85％,相对分析误差分别为2．96、1．98和1．04。结果表明,利用近红外光谱分析技术能够较准确地检测鱼粉中14种氨基酸;酪氨酸只能进行粗略估测;组氨酸、丝氨酸和半胱氨酸难于进行实际检测。     

放牧家畜养分消化率的测定

放牧家畜养分消化率反映牧草供给量和营养品质、家畜的健康状况以及生产性能等，是草原管理的关键指标之一。放牧家畜养分消化率的测定方法主要有直接测定和间接估测两种，直接法包括全收粪法、指示剂法、近红外光谱法（NIRS）等，间接法主要有体内或体外发酵法、牧草品质和气候等预测法。全收粪法是较精确的测定方法，然而耗时、费力，对家畜放牧行为有较大影响，难以体现牧场饲草供给的空间异质性；指示剂法根据不溶物的回收比例估测养分消化率，对家畜放牧行为干扰较小，指示剂不易收集，也不适于野性较高的放牧动物；NIRS法效率高、劳动强度低、成本小、不影响家畜，能够大尺度地估测放牧家畜的养分消化率，需要大量的实测数据完善预测模型；气候估测法快捷、省时，精确性较差，适于大时空尺度。放牧家畜的养分消化率与家畜、草地、放牧方法、温度、降水等生物因子，环境因子和社会因子密切相关，与遥感、无人机（UAV）和人工智能等结合，可以准确、快速地测定放牧家畜养分消化率，为草地生态修复与健康管理提供支撑。