大米直链淀粉含量的近红外光谱分析

大米的直链淀粉含量是影响大米蒸煮和加工特性的最重要因素之一,常被用作蒸煮米质构特性评价指标。该文对不同粒度、不同类型大米样品进行了近红外光谱分析,建立了大米直链淀粉含量的预测模型,（精米样品）预测值与化学分析值的相关系数达0.95。预测标准差、平均相对误差分别为0.56和3.1%。     

影响米粉品质的稻米质量指标分析

大米原料的质量是影响米粉加工性能、烹煮性能和感官品质的重要因素。影响米粉品质的大米原料质量指标主要有直链淀粉含量、胶稠度、糊化温度、蛋白质含量和储藏时间等,本文对近年来有关这些指标的研究进行了归纳总结,对各项指标对米粉品质的影响进行了逐一评述。通过总结分析可以得出,直链淀粉含量较高的籼米适宜于生产米粉;胶稠度与稻米其他各项指标有显著的相关性,可以作为选取米粉原料的主要指标;不同种类大米的糊化温度是不同的,可以根据糊化温度来指导米粉生产,避免淀粉过度糊化和米粉品质降低;蛋白质含量不仅决定大米的营养品质,也影响大米制品外观、加工和食用品质;使用陈化一段时间的大米制作米粉能够提升米粉品质。     

大米商业化辐照有关技术的研究

以杀虫为目的进行大米商业化辐照时，最小剂量为２００Ｇｙ，最大剂量为５００Ｇｙ，不均匀度为２．５。为保证辐照质量，应在大米加工后２０～３０ｄ内辐照。辐照前用包装袋密封。辐照后应防止外界害虫再次侵入。     

碱酶两步法制备大米蛋白的研究

该文采用碱法和酶法两步加工碎米，制备得到纯度较高的大米蛋白。通过研究米粉粒度、pH值、温度、水料比和时间对提取率的影响，确定碱法制备大米蛋白的较佳工艺条件为：米粉粒度80目，pH 11.0，温度50℃，水料比8，时间120 min。采用α-淀粉酶对大米蛋白进行纯化，酶解条件为：加酶量60 U/g，pH 6.0，温度50℃，时间30 min。在上述工艺条件下制备大米蛋白，提取率为73.22%，纯度为88.75%。     

大米胶稠度近红外光谱分析数学模型的建立

胶稠度是评价大米蒸煮食用品质的重要指标之一。研究了运用近红外光谱分析技术检测大米胶稠度的测试原理，对60个样品的光谱数据用偏最小二乘法(PLS)建立了测定大米胶稠度的数学模型，其回判结果与化学分析值之间的相关系数为0.95，建模标准差为0.66；用41个样品对建立的数学模型进行了交叉验证，其检测结果与用标准化学分析方法测得结果的相关系数达0.92，预测标准差为0.78。试验证明，可以利用近红外光谱分析技术对大米胶稠度进行快速检测。     

在MATLAB环境中基于计算机视觉技术的大米垩白检测

提出了利用计算机视觉系统代替人眼对垩白等大米质量参数进行自动检测的方法，以适应农业工程中的自动化的要求。为了验证检测方法的可行性，采用了MATLAB软件开发平台来构造基于计算机视觉的大米垩白检测算法。在对不同的大米图像处理的基础上，完成了对大米垩白参数垩白度和垩白粒率的测定。实验结果表明，本算法及程序设计是有效可靠的。     

大米直链淀粉分子量分布及分子旋转半径的研究

为了从大米淀粉中有效的分离出直链淀粉，建立直链淀粉分子量的分析方法，从而得出确切的大米直链淀粉分子结构方面的信息，通过使用流变仪控制大米淀粉的升温糊化过程，用浸出法在不同温度下分离出大米直链淀粉，利用高效液相分子排阻色谱(HPSEC)与多角度光散射仪(MALLS)及折光检测器(RI)连用系统，分析了大米直链淀粉的分子量分布和分子旋转半径。得到不同品种的大米淀粉中分离出的直链淀粉的重均分子量范围为3.29×10⁵～2.75×10⁶。研究表明，90℃以上高温不利于用浸出法从直链淀粉含量低的大米淀粉中分离直链淀粉，当温度低于90℃而高于糊化温度时，各个温度下均可用浸出法分离出直链淀粉，得到的直链淀粉具有十分相似的分子量分布、重均分子量和分子旋转半径。该方法是一种简便、快速的分离大米直链淀粉，并测定其分子量的方法，具有较高的准确性。     

挤压蒸煮大米啤酒辅料的糖化过程分析

为了合理地确定挤压蒸煮大米啤酒辅料的糖化工艺参数，通过挤压蒸煮大米不同糖化工艺的效果及传统蒸煮大米啤酒辅料糖化工艺的比较，分析了糖化过程。研究表明，两种大米啤酒辅料的糊化本质是相同的。不同的是，传统蒸煮大米啤酒辅料的双醪糖化工艺中的大米糊化过程中，包括了大米淀粉的糊化和液化过程；挤压蒸煮大米啤酒辅料仅完成了大米的糊化过程，其液化过程是在单醪糖化过程中完成。为此应调整挤压蒸煮大米的糖化工艺，充分发挥麦芽中的β-淀粉酶、α-淀粉酶及添加的耐高温α-淀粉酶的作用，使挤压蒸煮大米啤酒辅料的单醪糖化过程顺利进行，解决其糖化、过滤困难问题。     

基于计算机视觉的大米裂纹检测研究

针对人工目测的传统方法在进行米粒裂纹检验时存在主观性及随意性较大、效率较低、可重复性较差等缺点，在分析大米裂纹光学特征的基础上，在Visual C++ 6.0环境下开发了一套大米裂纹计算机识别系统，通过图像二值化、区域标记等方法从原始图像中提取单体米粒图像，并对提取出的单体米粒图像进行灰度拉伸变换处理以突出米粒裂纹特征，然后提取单体米粒的行灰度均值变化曲线，并对曲线进行加权滤波处理，提出了一种基于单体裂纹米粒图像行灰度均值变化特征的大米裂纹检测算法。运用该算法对从金优974、菲优600、冈优182、中优205、89-94等5类大米品种中各选取的6 组特殊类样品和5 组随机样品进行裂纹检测。试验结果表明，该系统对特殊类大米样品和随机大米样品裂纹率的判断准确率分别为98.37%和97.88%，为进一步完善大米品质的计算机视觉检测提供了理论和实践基础。     

基于近红外光谱和深度学习数据增强的大米品种检测

近红外光谱和深度学习结合的思路是大米品种检测的重要研究方向，其准确检测模型的建立依赖大规模的样本数据，然而采集和预处理样本耗时巨大，对准确性的提升造成限制。为解决上述不足，便于深入探究近红外光谱结合深度学习方法在大米品种检测领域应用的可行性，该研究提出基于近红外光谱结合改进型深度卷积生成式对抗神经网络（deep convolutional generative adversarial network，DCGAN）数据增强的大米品种检测方法。首先，在相同环境下采集4种大米品种的近红外光谱并对原始光谱数据进行预处理，使用去趋势校正（detrend correction，DC）和无信息变量消除算法（uninformative variable elimination，UVE）消除无用光谱特征点。然后，建立改进型DCGAN模型对预处理后的光谱数据进行数据增强，对比试验结果表明，改进型DCGAN相比与传统数据增强方法，改进型DCGAN生成数据的结构相似度指标更优。最后，研究不同数据增强方法结合不同分类方法建立大米品种分类模型的性能，对比试验结果表明，改进型DCGAN数据增强结合一维卷积神经网络（one-dimensional convolution neural network，1D-CNN）分类算法所建模型面向测试集的准确率最高，为98.21%，为简便准确的大米品种检测方案提供了新思路。     

Effect of Moisture Content at Harvest and Degree of Milling (Based on Surface Lipid Content) on the Texture Properties of Cooked Long-Grain Rice

The effects of the degree of milling (based on surface lipids content [SLC]) on cooked rice physicochemical properties were investigated. Head rice yield (HRY), protein, and SLC decreased with increasing milling, while the percent of bran removed and whiteness increased. Results showed that SLC significantly (P < 0.05) affected milled as well as cooked rice properties across cultivar, moisture content (MC) at harvest, and location (Stuttgart, AR, and Essex, MO). Cooked rice firmness ranged from 90.12 to 111.26 N after milling to various degrees (SLC). The decrease in cooked rice firmness with increasing milling was attributed to the lowering of total proteins and SLC. Cooked rice water uptake increased with increasing degree of milling. Water uptake by the kernel during cooking dictated the cooked rice firmness. The increase in cooked rice stickiness with increasing degree of milling was attributed to an increase in starch leaching during cooking because of the greater starch granule swelling associated with a greater water uptake.     

Degree of Milling Effects on Rice Pasting Properties

The effects of degree of milling on pasting properties of medium‐grain (cv. Bengal and Orion) and long‐grain rice (cv. Cypress and Kaybonnet) were quantified using a Brabender ViscoAmylograph and a Rapid Visco Analyser. For all the cultivars tested, surface and total lipid contents decreased as the degree of milling increased. The peak viscosities for all rice increased with the degree of milling and the rates of increase were higher for medium‐grain than long‐grain cultivars. Degree of milling did not have a consistent effect on final viscosity for all the cultivars tested.     

Effect of Milling on the Color,Nutritional Properties,and Antioxidant Contents of Glutinous Black Rice

Dehulled glutinous black rice (cv. Kam Doi Saket) was abrasively milled (0–60 s) to make a degree of milling (DOM) curve. The curve showed a nonlinear relationship between milling time and DOM. The slope sharply increased in the first 20 s; after this point, the increase was moderate. The redness of grain and flour surprisingly increased after milling, and the highest value was found at 20 s of milling. Protein, fat, and crude fiber were not uniformly distributed in the dehulled rice kernel. The rice retained only 76.95, 32.79, 20.24, and 36.57% of protein, fat, crude fiber, and ash, respectively, after milling for 60 s. Anthocyanins, γ‐oryzanol, and α‐tocopherol decreased by 74.49, 55.35, and 70.36%, respectively, after 10 s of milling. The scavenging activity decreased sharply when milling was carried out for longer than 30 s. The methanolic extract from rice milled for 60 s showed the lowest activity, which was 21.1 times lower than that of the dehulled rice extract. It was concluded that milling for 20 s was sufficient to remove most of the bran layer of the black rice sample, but 10 s of milling retained higher contents of nutritional components and rice antioxidants.     

Comparison of Three Extraction Systems for Determining Surface Lipid Content of Thickness‐Fractionated Milled Rice

The surface lipid content (SLC) of rice is often used to objectively measure the degree to which bran has been removed from rice kernels, commonly known as degree of milling (DOM). This study was conducted to evaluate new, rapid extraction technology for potential timesaving measurements of SLC of milled rice. The SLC of two long‐grain rice cultivars, Cypress and Drew, were determined using three extraction systems: Soxtec, accelerated solvent extraction (ASE), and supercritical fluid extraction (SFE). Before milling, rough rice was separated into three thickness fractions (<1.84, 1.84–1.98, and >1.98 mm) and samples from each thickness fraction were milled for durations of 10, 20, and 30 sec. Head rice collected from each milling duration was extracted using each of the three methods. Results showed that regardless of the extraction method, thinner kernels had lower SLC measurements than thicker fractions. In most cases, both the ASE and Soxtec produced SLC greater than that of the SFE. The ASE also showed SLC measurements at least as great as those from Soxtec extraction, suggesting that the ASE is as thorough in extracting lipids as commonly used methods.     

Effects of Postharvest Processing on Texture Profile Analysis of Cooked Rice

The effects of drying conditions, final moisture content, and degree of milling on the texture of cooked rice varieties, as measured by texture profile analysis, were investigated. Instrumentally measured textural properties were not significantly (α = 0.05) affected by drying conditions, with the exception of cohesiveness. Cohesiveness was lower in rice dried at lower temperatures (18°C or ambient) than in that dried at the higher commercial temperatures. Final moisture content and degree of milling significantly (α = 0.05) affected textural property values for adhesiveness, cohesiveness, hardness, and springiness; their effects were interdependent. The effects of deep milling were more pronounced in the rice dried to 15% moisture than that dried to 12%. In general, textural property values for hardness were higher and those for cohesiveness, adhesiveness, and springiness were lower in regular-milled rice dried to 15% moisture than in that dried to 12%. In contrast, hardness values were lower and cohesiveness, adhesiveness, and springiness values were higher in deep-milled rice dried to 15% moisture than in that dried to 12% moisture. Deep milling resulted in rice with lower hardness values and higher cohesiveness, adhesiveness, and springiness values.     

Use of microscopy to assess bran removal patterns in milled rice

During rice milling, the bran and germ are successively removed from the caryopsis (kernel). Because bran and germ contain large quantities of lipid, the amount of lipid remaining on the kernel surface may be used as a method for the assessment of milling quality. Bulk samples of rice pureline varieties and an experimental hybrid were milled for 0, 10, 20, 30, and 40 s. Scanning electron microscopy (SEM) revealed that brown rice kernels had large contours of linear protuberances and depressions running lengthwise along the kernel surface. The protuberances were abraded successively during milling, but varying amounts of material remained in the depressions. Light microscopy combined with the lipid-specific probes Nile Blue A or Sudan Black B demonstrated that the material in the depressions observed with SEM was lipid. Sections of whole, milled rice kernels, prepared using a modified sectioning technique and stained with Nile Blue A, showed that portions of the embryo remain after milling and that lipid is located on or near the surface of the kernel. Differences in quantity and distribution of residual lipid as milling duration increased were documented photographically to indicate the extent to which the bran and embryo components were removed during milling. This paper provides proof of concept that residual lipid is a robust measure of the degree of milling.     

Quality Characteristics of Long-Grain Rice Milled in Two Commercial Systems

Long-grain rice variety Kaybonnet was milled to three degree of milling (DOM) levels in two commercial milling systems (a single-break, friction milling system and a multibreak, abrasion and friction milling system) and separated into five thickness fractions. For both milling systems, the surface lipid content (SLC) and protein content of the milled rice varied significantly across kernel thickness fractions. SLC was influenced by DOM level more than by thickness, while the protein content was influenced by thickness more than by DOM level. Particularly at the low DOM levels, the thinnest kernel fraction (<1.49 mm) had higher SLC than the other kernel fractions. Protein content decreased with increasing kernel thickness to 1.69 mm, after which it remained constant. In both milling systems, thinner kernels were milled at a greater bran removal rate as indicated by SLC differences between the low and high DOM levels. For rice milled to a given DOM level, the multibreak system produced fewer brokens than did the single-break system.     

A Comparative Study Between the McGill #2 Laboratory Mill and Commercial Milling Systems

The degree of similarity between rice milled in a McGill #2 laboratory mill and commercial milling processes was evaluated using eight physical, physicochemical, and end‐use properties. There was no statistical difference between the two milling systems with respect to color parameters L* and a*, final viscosity, texture, and end‐use cooking properties (α = 0.05). Overall, the kernel dimensions of length, width, and thickness were less in the McGill #2 laboratory‐milled rice than the same rice milled commercially. The incidence of bran streaks and peak viscosity values were each higher when the rice sample was milled commercially in 27, and 28, respectively, of the 29 samples by means comparison. The decrease in kernel dimensions and incidence of bran streaks were attributed to the more aggressive nature of the single‐pass, batch milling system of the McGill #2 laboratory mill as compared with multipass, continuous milling systems that are used commercially. Finally, as surface lipid content (SLC) decreased, L* increased and a*, b*, and the incidence of bran streaks decreased for both milling systems.     

Evaluation of Operating Conditions for Surface Lipid Extraction from Rice Using a Soxtec System

The degree of milling (DOM) of rice is a measure of how well the germ and bran layers are removed from the surface of rice kernels during milling. Because the majority of rice kernel lipids are found on the surface, measuring the surface lipid content (SLC) of rice after milling may be one way to quantify the DOM of rice. While there are several methods to measure the lipid content (LC) of rice, there is not an established standard method for determining the SLC of milled rice. The objective of this study was to evaluate the primary operating variables of a Soxtec apparatus in measuring the SLC of milled rice. This was accomplished by varying the preextraction drying, boiling, rinsing, and postextraction drying durations, as well as the solvent used for extraction, to achieve the maximum extraction of lipids from rice. Experiments were performed on stored Oryza sativa L. ‘Cypress’ and ‘Bengal’ rice milled for 10, 30, and 60 sec. Results showed that durations of 1 hr of preextraction, 20 min of boiling, 30 min of rinsing, and 30 min of postextraction drying provided the maximum lipid extraction from milled head rice with petroleum ether. Of the three solvents tested, petroleum ether, and ethyl ether yielded similar extraction results.     

擦离式碾米机碾米室压力的研究

擦离式碾米机碾米室压力是影响成品米的含谷量和碎米率的关键因素。对N—70型擦离式碾米机的碾米室压力,通过单因素和多因素正交试验表明:碾米室压力呈周期性波动,其最大值比平均值大20％～35％;出米口轴向阻力以及碾米间隙对碾米室径向压力的影响均高度显著;擦离式碾米机用于加工籼稻,难以解决要求碎米率低和含谷量少之间的矛盾。