基于THz-TDS技术的豆科牧草种子的鉴定识别

发布时间：2018-01-15 05:13

本文关键词：基于THz-TDS技术的豆科牧草种子的鉴定识别　出处：《内蒙古大学学报(自然科学版)》2017年02期 　论文类型：期刊论文

【摘要】：利用太赫兹时域光谱(THz-TDS)技术对豆科牧草种子样品进行实验测试,得到了百脉根、柠条、毛苕籽、沙打旺和羊柴五种常见的牧草种子在太赫兹波段的时域谱,通过快速傅里叶变换得到频域谱,进而得到各牧草种子样品的吸收系数、折射率等光学参数.发现在0.2-1.4THz有效频域范围内,样品的吸收系数谱线以不同的斜率单调递增,样品的平均折射率在1.45-1.71之间,且彼此之间有较大的差异.同时利用偏最小二乘法回归(PLSR)分别对五种豆科牧草的35个样本的吸收系数和折射率做了模型的建立和统计分析.结果表明,利用折射率建立的PLS模型预测效果最好,其相关系数(R)和均方根误差(RMSE)分别为0.9983和0.3157,更能准确鉴定未知样品的种类.研究表明,太赫兹时域光谱技术结合统计分析方法可以作为无损鉴定牧草种子真伪和识别牧草种子种类的一种有效手段,不仅丰富了牧草种子的检测手段,而且有望发展建立标准检测的数据库,最终实现快速、批量、经济鉴定识别牧草种子的目的.
[Abstract]:Using terahertz time-domain spectroscopy (THz-TDS) experiments on legume seed samples, the Lotus corniculatus, Caragana, Vicia villosa seeds, temporal and Hedysarum five kinds of Astragalus adsurgens grass seed in the terahertz spectrum, frequency spectrum obtained by fast Fourier transform, and then obtain the absorption coefficient of each sample of forage seed, optical parameters the refractive index was found effective in the frequency domain. The range of 0.2-1.4THz, the absorption coefficient of the sample line increased in different slope monotonously, the average refractive index of the sample at 1.45-1.71, and there are great differences between each other. At the same time using partial least squares regression (PLSR) respectively. The absorption coefficient and refractive index of 35 samples of five species of Leguminosae the forage establishment and statistical analysis model. The results show that the PLS model is used to establish the best prediction effect of refractive index, the correlation coefficient (R) and the root mean square error (RMSE) Were 0.9983 and 0.3157, more accurate species identification of unknown samples. The results show that the terahertz time-domain spectroscopy combined with statistical analysis method can be used as an effective means of nondestructive identification of forage seed authenticity and species identification of forage seeds, not only enrich the detection methods of forage seed, and is expected to establish a standard testing database, realize fast, batch, economic identification of forage seed purpose.

【作者单位】：油气光学探测技术北京市重点实验室;全国石油与化工行业油气太赫兹波谱与光电检测重点实验室;中国石油大学(北京)光传感与光探测实验室;内蒙古自治区草原工作站;
【基金】：国家自然科学基金(项目批准号:61405259)
【分类号】：S54;O433
【正文快照】： 牧草是指供饲养的牲畜使用的草或其他草本植物,牧草再生力强,且富含蛋白质、脂肪、纤维和各种微量元素,成为饲养家畜的首选.牧草品种的优劣直接影响到畜牧业经济效益的高低,因此牧草品种的鉴定是草业科学研究中的一项基本内容[1].目前,已有的牧草品种鉴别可分为现场鉴别和实验

【相似文献】