基于LS-SVM的烤煙煙葉產(chǎn)地判別
發(fā)布時(shí)間:2019-08-24 來源: 人生感悟 點(diǎn)擊:
摘要:為了探索一種快速有效的烤煙煙葉產(chǎn)地鑒別方法,利用近紅外光譜技術(shù)結(jié)合最小二乘支持向量機(jī)(LS-SVM)對(duì)烤煙煙葉的產(chǎn)地進(jìn)行了判別。選擇云南、湖北、河南三地不同等級(jí)烤煙煙葉作為研究對(duì)象,對(duì)原始光譜數(shù)據(jù)進(jìn)行平滑和附加散射校正(MSC)預(yù)處理后再進(jìn)行主成分分析,選擇4~12個(gè)主成分作為輸入變量進(jìn)行LS-SVM建模。結(jié)果顯示,該LS-SVM模型預(yù)測(cè)效果較好,預(yù)測(cè)相關(guān)系數(shù)rp≥0.990 7,預(yù)測(cè)標(biāo)準(zhǔn)誤差(SEP)和預(yù)測(cè)均方根誤差(RMSEP)分別為1.755 1和1.737 3,優(yōu)于偏最小二乘回歸(PLS)的預(yù)測(cè)結(jié)果,基于LS-SVM的近紅外光譜技術(shù)能夠很好地對(duì)煙葉產(chǎn)地進(jìn)行判別。
關(guān)鍵詞:煙葉;產(chǎn)地判別;近紅外光譜;最小二乘支持向量機(jī)
中圖分類號(hào):TN219 文獻(xiàn)標(biāo)識(shí)碼:A 文章編號(hào):0439-8114(2012)03-0583-03
。桑洌澹睿簦椋妫椋悖幔簦椋铮 of Producing Area of Tobacco Leaf Based on LS-SVM
。冢龋粒危 Ying1a,1b,HE Li-yuan1b,YE Ying-ze1c,WU Zhao-hui2
。ǎ幔 College of Science; b. College of Resources and Environment; c. Network Center, 1.Huazhong Agricultural University, Wuhan 430070, China; 2. Tobacco Research Center of Henan Academy of Agricultural Sciences, Xuchang 461000, Henan, China)
。粒猓螅簦颍幔悖簦 In order to explore a fast and efficient method which determines the producing area of tobacco leaf, near-infrared reflectance spectroscopy with least squares-support vector machines (LS-SVM) was applied to determine producing area of tobacco leaf. Three producing areas including Yunnan, Hubei and Henan were selected as the research objects. As the pretreatments of the optimal smoothing way, moving average with three segments and multiplication scatter correction (MSC) were applied to reduce the noise of the spectra. After the principle component analysis, 4 to 12 principal components (PCs) were chosen as the inputs of LS-SVM models. The Results show that the prediction performance of the LS-SVM model with 12 PCs is better than partial least square(PLS) model. Its correlation coefficient of prediction set (rp) is 0.990 7, standard error of prediction (SEP) is 1.755 1, and root mean square error of prediction (RMSEP) is 1.737 3. It is concluded that NIR spectroscopy with LS-SVM is a feasible method to determine the producing area of tobacco leaf.
。耍澹 words: tobacco leaf; origin discriminant; NIR spectroscopy; least squares-support vector machines (LS-SVM)
煙草是我國重要的經(jīng)濟(jì)作物,煙葉的品質(zhì)與遺傳因素、栽培措施、調(diào)制技術(shù)和產(chǎn)地環(huán)境等密切相關(guān)。其中,產(chǎn)地環(huán)境(海拔、溫濕度、氣候條件等)對(duì)煙葉品質(zhì)的影響極為明顯,也是構(gòu)成不同品牌卷煙特有風(fēng)格的基礎(chǔ),但不同產(chǎn)地的煙葉特征迄今難以量化描述。目前,對(duì)烤煙煙葉產(chǎn)地的判別除依賴感官評(píng)定外,需要對(duì)其化學(xué)成分進(jìn)行分析,判別過程費(fèi)時(shí)、費(fèi)力。因此,研究一種能夠快速、準(zhǔn)確地對(duì)烤煙煙葉產(chǎn)地進(jìn)行判別的方法非常必要。
。停幔瑁岬龋郏保莶捎蒙窠(jīng)元網(wǎng)絡(luò)方法對(duì)美國本土及國外1 000多個(gè)煙葉樣品的近紅外光譜(NIRs)信息進(jìn)行分析,對(duì)本國煙葉取得了很好的模式識(shí)別結(jié)果。國內(nèi)研究人員曾采用NIR法預(yù)測(cè)了煙草根、莖、葉中的蛋白質(zhì)、總糖、總氮、總植物堿等[2,3],采用主成分分析的馬氏距離法判別煙葉產(chǎn)地歸屬,獲得了較佳的識(shí)別準(zhǔn)確率[4]。但上述研究均需要對(duì)煙葉進(jìn)行切絲過篩,屬于有損檢測(cè)且費(fèi)時(shí)費(fèi)力。用NIR法專門針對(duì)收購環(huán)節(jié)進(jìn)行完整煙葉品質(zhì)分析預(yù)測(cè)煙葉產(chǎn)地尚無研究報(bào)道。試驗(yàn)采用近紅外波段(867~258 9 nm)進(jìn)行光譜掃描,應(yīng)用最小二乘支持向量機(jī)(LS-SVM),建立了LS-SVM判別分析組合模型,實(shí)現(xiàn)了烤煙煙葉產(chǎn)地的快速準(zhǔn)確判別。
。 材料與方法
。保 儀器及參數(shù)
試驗(yàn)使用光譜檢測(cè)設(shè)備是Ocean Optics公司的NIR256-2.5光纖光譜儀,配套的QBIF600-VIS-BX白金級(jí)Y形分叉光導(dǎo)纖維探頭,儀器光譜采樣間隔6 nm,測(cè)定波長867~2 589 nm,光纖探測(cè)器與樣品垂直,暗室溫度18~22 ℃,相對(duì)濕度22%~25%,以14.5 V、50W鹵素?zé)魹槲┮还庠,光源與樣品夾角45°。開機(jī)預(yù)熱1 h后進(jìn)行光譜掃描,采樣方式是漫反射,采樣軟件是機(jī)器自帶的Spectra Suite。分析軟件采用ASD View Spec Pro、Unscramble V9和DPS(Data Procession System for Practical Statistics)。積分時(shí)間設(shè)置為250 mm,平滑度設(shè)置為9,平均次數(shù)為3,即對(duì)每個(gè)樣品自動(dòng)掃描3次取平均值。
。保 樣本制備
收集了2010年10月云南、河南、湖北三省的煙草公司提供的已由專家人工定級(jí)的煙葉。為保證試驗(yàn)結(jié)果的代表性,每個(gè)產(chǎn)地?zé)熑~按7個(gè)分組每組1~4個(gè)等級(jí)隨機(jī)選擇90個(gè)樣本。根據(jù)文獻(xiàn)[5]報(bào)道,直接將煙葉樣品平鋪置于載物臺(tái)上,采用漫反射模式采集近紅外光譜,光譜掃描穩(wěn)定后進(jìn)行數(shù)據(jù)采集。保存3條光譜曲線,以其平均光譜作為最終的反射光譜。從全部270個(gè)樣本中,每個(gè)產(chǎn)地隨機(jī)選擇30個(gè)共90個(gè)樣本作為預(yù)測(cè)集,剩余的180個(gè)樣本作為建模集。
相關(guān)熱詞搜索:烤煙 判別 煙葉 產(chǎn)地 LS
熱點(diǎn)文章閱讀