本文選題：玉米　切入點(diǎn)：近紅外光譜　出處：《光譜學(xué)與光譜分析》2017年05期

【摘要】：以2009年產(chǎn)自海南的農(nóng)華101玉米種子作為研究對(duì)象,基于近紅外漫透射光譜法(波長(zhǎng)范圍908.1~1 677.2nm),研究了一種高強(qiáng)高效近紅外光源在不同光源電壓以及光源至光譜儀不同距離兩種條件下對(duì)玉米雜交種鑒別的影響。對(duì)光譜進(jìn)行一階導(dǎo)數(shù)、矢量歸一化的預(yù)處理后,使用主成分分析(PCA)和正交線性判別分析(OLDA)提取光譜特征,使用支持向量機(jī)(SVM)分別建立種子純度鑒定模型,統(tǒng)計(jì)不同實(shí)驗(yàn)條件下的識(shí)別率。結(jié)果表明,在電壓較低或者光源至光譜儀的距離較大的時(shí)候,光源強(qiáng)度較低,得到的光譜曲線有較多的毛刺,此時(shí)的識(shí)別率較低,增大電壓或者降低光源至光譜儀的距離時(shí),光譜曲線變得較為平滑,識(shí)別率明顯升高,說(shuō)明在一定范圍內(nèi)增大光源強(qiáng)度會(huì)提高模型的正確鑒定率。
[Abstract]:In the 2009 annual Hainan Nonghua 101 corn seeds as the research object, near infrared transmission spectroscopy (based on the wavelength range of 908.1~1 677.2nm), studied a kind of high efficient near infrared light source effects on maize hybrid identification in different light source voltage and source to the spectrometer in different distance under two conditions. The first derivative the spectral vector normalization preprocessing, using principal component analysis (PCA) and orthogonal linear discriminant analysis (OLDA) to extract spectral feature, using support vector machine (SVM) were established model for seed purity identification, the recognition rate under different experimental conditions. The results show that the voltage is low or the light source and a spectrometer the distance is larger when the light intensity is low, the spectral curves have more burr, the low recognition rate, increase or reduce the voltage source to the spectrometer distance, spectral curve It is more smooth and the recognition rate increases obviously. It shows that increasing the light intensity within a certain range will improve the correct identification rate of the model.

【作者單位】： 中國(guó)農(nóng)業(yè)大學(xué)信息與電氣工程學(xué)院;農(nóng)業(yè)部農(nóng)業(yè)信息獲取技術(shù)重點(diǎn)實(shí)驗(yàn)室;
【基金】：大北農(nóng)青年學(xué)者研究計(jì)劃項(xiàng)目(1081-2413001)資助
【分類(lèi)號(hào)】：O657.33;S513
，

本文編號(hào)：1674550