基于近红外光谱的单籽粒水稻种子品质检测的方法研究
发布时间：2018-08-13 15:34
【摘要】：近红外光谱(Near infrared spectroscopy,NIRS)技术因具有快速、无损等优点已经被广泛的用于定性、定量分析单粒谷物种子,这种分析方式不仅克服了传统检测方法中耗时长、工作量较大、不环保等缺点,且能够更加精确的获取具有目标性状的种子,有利于种子的品质分析、收购、储藏等环节的实时监管。水稻是全球重要的粮食作物之一,由于存在粒径小、种壳干扰等原因,其在近红外光谱技术的分析应用方面的研究较少。本论文基于这一研究现状,细致分析了单粒糙米和单粒水稻种子的近红外光谱特点,并成功建立相应的分析模型,证明了近红外光谱技术能够用于单粒糙米、稻种的无损品质分析。获得的主要结果如下:1)建立单粒糙米水分含量的近红外漫反射光谱模型,结果表明:采用5292～5616cm-1、7236～7600cm-1、7884-8208cm-1波数范围,用标准正态变换光谱预处理(SNV)建立的单粒糙米水分偏最小二乘(PLS)模型的预测能力最佳,其决定系数(R2)为0.98,预测误差均方根(RMSEP)为1.01%;采用8285.12cm-1、7158.84cm-1、5492.56cm-1这3个波数变量建立的单粒糙米含水量多元线性回归(MLR)模型,变量最少且预测能力较优,其R2为0.9661,RMSEP为1.137%。2)分析了糙米和水稻种子的近红外光谱特征,并细致对比了不同光谱采集方式和不同光谱预处理下单粒糙米和单粒种子蛋白质模型。单粒糙米的蛋白质定量分析中,透射和透反射光谱采集方式下均能建立较好的近红外光谱蛋白质模型;其中在透反射光谱采集方式下,选择4000～9000cm-1波数范围,SNV光谱预处理下,单粒糙米的蛋白质模型的R2为0.941,RMSEP为0.338%。单粒水稻种子光谱对其糙米蛋白质含量的定量分析中,透反射光谱能够增强光谱的信噪比,选择6500～9100cm-1波数范围和SNV光谱预处理,其模型的RMSEP为0.806%,具有一定的相关性;而透射光谱采集方式下,选择7200～9100cm-1波数范围和SNV光谱预处理,单粒种子光谱和其糙米蛋白质含量的相关性极好,模型的R2为0.964,RMSEP为0.244%。这一结果表明近红外光谱技术能够用于单粒糙米和单粒水稻种子的无损定量分析。3)研究了低温等离子体对糙米生长活力的影响,并采用主成分分析方法对处理后的糙米近红外光谱进行初步定性分析。结果显示:空气气氛下等离子体对糙米生长初期活力的促进效果要优于氮气气氛,空气气氛下最佳处理条件为360W-5分钟,氮气气氛下最佳处理条件为360W-10分钟,PCA结果证明经等离子体处理后糙米的近红外光谱与未处理糙米的光谱之间存在差异,说明等离子体处理改变了糙米的物理化学性状,且近红外光谱技术具有潜力应用于初期评估等离子体对糙米的作用及效果。
[Abstract]:Near-infrared spectroscopy (Near infrared) technique has been widely used in qualitative and quantitative analysis of single grain seeds because of its advantages of fast and nondestructive. This method not only overcomes the time-consuming and heavy workload of traditional detection methods, but also has the following advantages: (1) Near-infrared spectroscopy (NIR) has been widely used for qualitative and quantitative analysis of single grain seeds. Not environmental protection and other shortcomings, and more accurate access to seeds with target traits, conducive to seed quality analysis, acquisition, storage and other links of real-time supervision. Rice is one of the most important food crops in the world. Because of the small particle size and the interference of seed shell, there is little research on the analysis and application of near infrared spectroscopy (NIR). Based on the present research situation, the characteristics of NIR spectra of single brown rice and single grain rice seed were analyzed in detail, and the corresponding analysis model was established successfully, which proved that NIR spectroscopy technology could be used in single brown rice. Non-destructive quality analysis of rice seeds. The main results obtained are as follows: (1) the near infrared diffuse reflectance spectral model of water content of single brown rice is established. The results show that the wave number range of 7884-8208 cm-1 is 7884-8208 cm ~ (-1) using 5292 ~ 1 ~ 5616 cm ~ (-1) ~ (-1) ~ 7236n ~ 7600cm ~ (-1) and 7884-8208 cm ~ (-1). The partial least square (PLS) model of water content in single brown rice pretreated with standard normal transform spectrum (SNV) has the best prediction ability. The coefficient of determination (R2) is 0.98, the root mean square (RMSEP) of prediction error is 1.01.The multivariate linear regression (MLR) model of water content of single brown rice is established by using the three wavenumber variables of 8285.12cm-1a 7158.84cm-1n-1 5492.56 cm ~ (-1). The RMSEP (R2 = 0.9661) was used to analyze the near infrared spectra of brown rice and rice seeds, and the models of order grain brown rice and single seed protein were compared in detail. In protein quantitative analysis of single brown rice, a better protein model of near infrared spectrum can be established under transmission and transmission spectrum acquisition, in which the 4000~9000cm-1 wave number range is chosen to be pretreated with SNV spectrum. The R2 of single brown rice protein model was 0.941 and RMSEP was 0.3338. In the quantitative analysis of protein content in brown rice by single grain rice seed spectrum, the transmittance reflectance spectrum can enhance the signal-to-noise ratio of the spectrum. The range of 6500~9100cm-1 wave number and the pretreatment of SNV spectrum are selected. The RMSEP of the model is 0.806, which has certain correlation. However, the correlation between the single seed spectrum and the protein content of brown rice was excellent when the range of 7200~9100cm-1 wave number and the pretreatment of SNV spectrum were selected. The R2 of the model was 0.964RMSEP 0.244. The results showed that Near-infrared spectroscopy could be used for quantitative analysis of single brown rice and rice seeds. The effects of low temperature plasma on the growth activity of brown rice were studied. The method of principal component analysis (PCA) was used to analyze the NIR spectra of brown rice. The results showed that the effect of plasma on the initial vigor of brown rice was better than that in nitrogen atmosphere, and the best treatment condition was 360W-5 minutes. The optimum treatment conditions in nitrogen atmosphere were as follows: the results of 360W-10 showed that there were differences between the near infrared spectra of brown rice treated by plasma and those of untreated brown rice, which indicated that plasma treatment changed the physical and chemical properties of brown rice. Near-infrared spectroscopy has the potential to be used to evaluate the effect of plasma on brown rice.
【学位授予单位】：中国科学技术大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：S511
本文编号：2181410
[Abstract]:Near-infrared spectroscopy (Near infrared) technique has been widely used in qualitative and quantitative analysis of single grain seeds because of its advantages of fast and nondestructive. This method not only overcomes the time-consuming and heavy workload of traditional detection methods, but also has the following advantages: (1) Near-infrared spectroscopy (NIR) has been widely used for qualitative and quantitative analysis of single grain seeds. Not environmental protection and other shortcomings, and more accurate access to seeds with target traits, conducive to seed quality analysis, acquisition, storage and other links of real-time supervision. Rice is one of the most important food crops in the world. Because of the small particle size and the interference of seed shell, there is little research on the analysis and application of near infrared spectroscopy (NIR). Based on the present research situation, the characteristics of NIR spectra of single brown rice and single grain rice seed were analyzed in detail, and the corresponding analysis model was established successfully, which proved that NIR spectroscopy technology could be used in single brown rice. Non-destructive quality analysis of rice seeds. The main results obtained are as follows: (1) the near infrared diffuse reflectance spectral model of water content of single brown rice is established. The results show that the wave number range of 7884-8208 cm-1 is 7884-8208 cm ~ (-1) using 5292 ~ 1 ~ 5616 cm ~ (-1) ~ (-1) ~ 7236n ~ 7600cm ~ (-1) and 7884-8208 cm ~ (-1). The partial least square (PLS) model of water content in single brown rice pretreated with standard normal transform spectrum (SNV) has the best prediction ability. The coefficient of determination (R2) is 0.98, the root mean square (RMSEP) of prediction error is 1.01.The multivariate linear regression (MLR) model of water content of single brown rice is established by using the three wavenumber variables of 8285.12cm-1a 7158.84cm-1n-1 5492.56 cm ~ (-1). The RMSEP (R2 = 0.9661) was used to analyze the near infrared spectra of brown rice and rice seeds, and the models of order grain brown rice and single seed protein were compared in detail. In protein quantitative analysis of single brown rice, a better protein model of near infrared spectrum can be established under transmission and transmission spectrum acquisition, in which the 4000~9000cm-1 wave number range is chosen to be pretreated with SNV spectrum. The R2 of single brown rice protein model was 0.941 and RMSEP was 0.3338. In the quantitative analysis of protein content in brown rice by single grain rice seed spectrum, the transmittance reflectance spectrum can enhance the signal-to-noise ratio of the spectrum. The range of 6500~9100cm-1 wave number and the pretreatment of SNV spectrum are selected. The RMSEP of the model is 0.806, which has certain correlation. However, the correlation between the single seed spectrum and the protein content of brown rice was excellent when the range of 7200~9100cm-1 wave number and the pretreatment of SNV spectrum were selected. The R2 of the model was 0.964RMSEP 0.244. The results showed that Near-infrared spectroscopy could be used for quantitative analysis of single brown rice and rice seeds. The effects of low temperature plasma on the growth activity of brown rice were studied. The method of principal component analysis (PCA) was used to analyze the NIR spectra of brown rice. The results showed that the effect of plasma on the initial vigor of brown rice was better than that in nitrogen atmosphere, and the best treatment condition was 360W-5 minutes. The optimum treatment conditions in nitrogen atmosphere were as follows: the results of 360W-10 showed that there were differences between the near infrared spectra of brown rice treated by plasma and those of untreated brown rice, which indicated that plasma treatment changed the physical and chemical properties of brown rice. Near-infrared spectroscopy has the potential to be used to evaluate the effect of plasma on brown rice.
【学位授予单位】：中国科学技术大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：S511
【参考文献】
相关期刊论文 前9条
1 郝光宗,邢丽缘,梁强威;饱和盐水溶液湿度固定点(1)——原理及制备[J];传感器世界;1999年11期
2 符文英,陈俊;稻米营养品质研究综述[J];海南大学学报(自然科学版);1997年01期
3 褚小立,袁洪福,陆婉珍;光谱多元校正中的模型传递[J];光谱学与光谱分析;2001年06期
4 李天真;稻米碾米品质及相关因素影响的研究进展[J];浙江农业科学;2005年03期
5 胡孔峰;杨泽敏;雷振山;;中国稻米品质研究的现状与展望[J];中国农学通报;2006年01期
6 邹小波;赵杰文;;用遗传算法快速提取近红外光谱特征区域和特征波长[J];光学学报;2007年07期
7 张海艳;;低温对鲜食玉米种子萌发及幼苗生长的影响[J];植物生理学报;2013年04期
8 宋涛;张凤枰;刘耀敏;吴宗文;索有瑞;;透反射近红外光谱法快速测定大豆油中的脂肪酸[J];光谱学与光谱分析;2012年08期
9 Pei Wang;Zhiguo Yu;;Species authentication and geographical origin discrimination of herbal medicines by near infrared spectroscopy:A review[J];Journal of Pharmaceutical Analysis;2015年05期
相关硕士学位论文 前1条
1 马晓娟;关于稻米的蒸煮及食味评价的研究[D];扬州大学;2005年
，本文编号：2181410
本文链接：https://www.wllwen.com/shoufeilunwen/zaizhiyanjiusheng/2181410.html
