CN103424542A - Tobacco leaf quality evaluation method based on sensory evaluation - Google Patents

Abstract

The invention discloses a tobacco leaf quality evaluation method based on sensory evaluation. The tobacco leaf quality evaluation method is based on tobacco leaf sensory evaluation, various algorithms are used for conducting preliminary evaluation and selection on indicators of obtained samples, main physicochemical indexes which influence tobacco leaf characteristics are initially determined, tobacco leaf sensory evaluation quality is divided in an ordered sample Fisher classification or hierarchical clustering method, samples to which divisional subsections belong are analyzed and evaluated, then the main physicochemical indexes which influence the tobacco leaf characteristics are determined, the optimal sample is determined at the same time, the suitable range of the main physicochemical indexes which influence tobacco leaf sensory evaluation is obtained according to the limiting condition that the number of samples is larger than 60% of conformity, tobacco industrial enterprises are guided to conduct tobacco leaf raw material purchase according to the main physicochemical indexes and the suitable range of the physicochemical indexes, tobacco leaf samples meeting requirements are searched in the reverse direction according to the suitable range of the physicochemical indexes, then the growth ecological environment, variety, cultivation techniques and matched production measures of the samples are obtained, and tobacco commercial enterprises are guided to conduct tobacco production.

Description

A kind of quality of tobacco evaluation method based on the sensory evaluating smoking

Technical field

The present invention relates to the Evaluation for Tobacco Quality research field, specifically belong to a kind of Quality of tobacco evaluation method based on the sensory evaluating smoking.

Background technology

Tobacco is most important industrial crops, is the important component part of national economy, in agricultural and trading, occupies critical role.Quality of tobacco is a comprehensive and dynamic concept, is also a relative term.5 aspects such as quality of tobacco comprises presentation quality, the quality of smokeing panel test, physical behavior, chemical composition and security.Sound tobacco and tobacco product have the chemical composition, nontoxic comparatively safe of perfect external appearance characteristic, good interior quality (being fragrance and jealous), perfect physical behavior, coordination.The quality of tobacco evaluation is the basis of numerous tobaccos forward position research direction, what emphasize is the actual application value of tobacco leaf to tobacco industrial enterprise, and the quality of tobacco evaluation result can be described as the tobacco leaf usability evaluation in some sense to actual guiding value and the meaning of tobacco industrial enterprise production of cigarettes and tobacco commercial enterprise leaf tobacco production.

At present, quality of tobacco evaluation study content is more, but many and relations between any two single from quality of tobacco utilize mathematical statistics method to carry out assay, or from experience in the past, quality of tobacco is carried out to artificial suitable division and qualitative evaluation, though the formation of result has embodied the control to raw tobacco material quality stability, degree of variation to a certain extent, but, to instructing the raw tobacco material buying of cigarette product formula demand, the raising of guiding base raw tobacco material quality level and the outstanding power of work of style characteristic are not strong.

Summary of the invention

The technical problem to be solved in the present invention is to provide a kind ofly can improve understanding to the quality of tobacco feature better, find existing problems, for the buying of tobacco industrial enterprise raw tobacco material and the improvement of tobacco commercial enterprise flue-cured tobacco base production technology provide decision-making foundation Quality of tobacco evaluation method based on the sensory evaluating smoking, can overcome the deficiencies in the prior art.

The present invention realizes by following technical proposals: a kind of quality of tobacco evaluation method based on the sensory evaluating smoking, with all improved final purpose of tobacco be all using the tobacco leaf sensory evaluating smoking well as target as starting point, take the tobacco leaf sensory evaluating smoking as basis, utilize simple correlation, partial correlation and successive Regression, the Double Selection successive Regression, path analysis, and the method such as artificial neural network is carried out preliminary assessment and screening to the indices that obtains sample, the main physical and chemical index that affects the tobacco leaf characteristic is tentatively determined, utilize ordered sample Fisher classification or Hierarchical Clustering method to be divided tobacco leaf sensory evaluating smoking quality, utilize the test of hypothesis of multivariate analysis of variance single-factor and multiple ratio to carry out assay in sample under the segmentation of cutting apart and then determine the main physical and chemical index that affects the tobacco leaf characteristic, carry out determining of optimum sample simultaneously, and according to sample number>60% degree of conformity is restrictive condition, obtain and affect the main physical and chemical index optimum range of tobacco leaf sensory evaluating smoking, and instruct the buying of tobacco industrial enterprise raw tobacco material with utilization and according to main physical and chemical index optimum range according to main physical and chemical index and optimum range thereof, the tobacco leaf sample that reverse find meets the demands, continuous and draw the ecologic environment of this sample growth, kind and cultivation technique and the measure that forms a complete production network, instruct tobacco commercial enterprise leaf tobacco production.

Above-mentioned Quality of tobacco evaluation method based on the sensory evaluating smoking is, comprise the following steps:

(1) sample acquisition, require institute's collecting sample and grade contain wide, purity is high, representativeness is strong;

(2) pattern detection, obtain sample data by sensory evaluating smoking, industrial applicability analysis, mensuration presentation quality, physical characteristics and all kinds of chemical compositions of profile sample and the relevant index of quality;

(3) former data are processed, and step (2) the data obtained resource is arranged, and make it to be beneficial to statistical study, extrapolate sugared alkali ratio, two sugar ratios, potassium chlorine ratio, two sugar difference and all kinds of indexs that affect cigarette quality, the rejecting abnormalities value;

(4) analyzed according to the actual conditions of step (3) gained sample data and sample sampling spot, guaranteed that sample data meets representativeness, wide in range property and discreteness;

(5) evaluation index primary election, because all improved final purpose of tobacco is all that to take the tobacco leaf sensory evaluating smoking be well target, so take the tobacco leaf sensory evaluating smoking as basis, utilize simple correlation, partial correlation and successive Regression, Double Selection successive Regression, path analysis, and the method such as artificial neural network carries out preliminary assessment and screening to the quality evaluation index in step (3), preliminary screening goes out to affect the main physical and chemical index of tobacco leaf characteristic;

(6), on the basis of step (5), utilize ordered sample Fisher classification or Hierarchical Clustering method to the smoked panel test division of quality of sample; 1) the affiliated sample of the segmentation of cutting apart is utilized to the test of hypothesis of multivariate analysis of variance single-factor and multiple ratio, further the screening and assessment index, determine relatively convenient, accurate and can reflect tobacco style and quality stability control index; 2) sample decomposition according to methods such as ordered sample Fisher classification or Hierarchical Clusterings, tobacco leaf sensory evaluating smoking quality carried out, in conjunction with central value ± SD(standard deviation), the confidence level that the numerical value interval of limiting is dropped in population sample is 68.2%, central value ± 1.96SD(standard deviation) be 95%, central value ± 2.57SD(standard deviation) be 99% requirement, determine optimum sample sensory evaluating smoking scope, then determine optimum sample cluster;

(7) verification and debugging optimum range

According to sample number > 60% degree of conformity is restrictive condition, in conjunction with cigarette enterprise to the raw tobacco material quality requirements, introduce the external certificate sample, carry out the degree of conformity checking according to determined main physical and chemical index in step (6), and then debugging and revision index of correlation parameter, obtain the main physical and chemical index optimum range of tobacco leaf;

(8) instruct the buying of tobacco industrial enterprise raw tobacco material and utilize

The main physical and chemical index of tobacco leaf and optimum range thereof according to obtaining in step (7), carry out tobacco leaf classification and purchase, instructs the buying of tobacco industrial enterprise raw tobacco material and utilize;

(9) reckoning of ideotype growing way appearance and the measure that forms a complete production network

According to the main physical and chemical index of tobacco leaf and the optimum range thereof that obtain in step (7), the tobacco leaf sample that reverse find meets the demands, continue and draw ecologic environment, kind and cultivation technique and the measure that forms a complete production network that this sample is grown, instructing tobacco commercial enterprise leaf tobacco production;

Aforesaid Quality of tobacco evaluation method based on the sensory evaluating smoking is, in step (1), sample divides typical sample and outside two parts of checking sample of introducing, and when obtaining sample, correspondence is carried out investigation the record of sampling spot ecology, kind and cultivation technique:

The characteristics such as that the selected grade that requires a, typical sample contains is wide, purity is high, representative strong, this part sample detects the parameter of obtaining will be as basic assay data;

B, the outside checking sample of introducing, require originate wide, wide ranges, can represent that Cigarette Industrial Enterprise raw tobacco material buying level gets final product, and this part sample detects the parameter of obtaining will or revise data as checking.

Compared with the prior art, the present invention serves as theme with the tobacco leaf quality of smokeing panel test, adopt a large amount of mathematical statistics methods, by tobacco leaf presentation quality, physical characteristics, main chemical compositions content, the quality of smokeing panel test are carried out to comprehensive A+E, filter out the key index that affects quality of tobacco, and obtain the suitable threshold value of the main physical and chemical index of tobacco leaf.Can realize improving better understanding to the quality of tobacco feature, find existing problems, for the buying of tobacco industrial enterprise raw tobacco material and the improvement of tobacco commercial enterprise flue-cured tobacco base production technology provide decision-making foundation.

Its advantage is: (1) take the tobacco leaf sensory evaluating smoking as basis, and tobacco leaf presentation quality, physical characteristics, main chemical compositions content, the quality of smokeing panel test have been carried out to system and comprehensive A+E; (2) serve as theme with the tobacco leaf quality of smokeing panel test, take full advantage of mathematical statistics method, carry out the assay of science, avoided the impact of human factor on evaluation result; (3) can filter out by evaluation method and system the main physical and chemical index that affects the tobacco style characteristic, simultaneously, obtain main physical and chemical index optimum range, and realize checking and the revision of optimum range; (4) screening of evaluation index and optimum range determines, the Continuous optimization and the correction that can be quality of tobacco system index provide approach, improve the understanding of people to the quality of tobacco feature, for the buying of tobacco industrial enterprise raw tobacco material and utilization and the improvement of tobacco commercial enterprise flue-cured tobacco base production technology provide decision-making foundation; (5) application of this method and evaluation system can propose work, desired cigarette strain ideotype portrait and the measure that forms a complete production network of commercial business's industry, instructs the buying of tobacco industrial enterprise raw tobacco material and tobacco commercial enterprise leaf tobacco production; (6) quality of this evaluation method and system evaluation tobacco leaf has stronger objectivity, systematicness and specific aim, and operation instruction preferably and practicality; (7) this evaluation method and system arise from production, stop in production, realize preferably combination and the circulation of product-pin-product.

The accompanying drawing explanation

Fig. 1 is logic diagram of the present invention.

Embodiment

Embodiment 1: How the present invention is directed to people utilizes mathematical statistics method or from experience, quality of tobacco is carried out the deficiency of assay from relation single and between any two, and evaluation result is to instructing the raw tobacco material buying of cigarette product formula demand, the problem such as the raising of guiding base raw tobacco material quality level and the outstanding power of work of style characteristic are not strong.Proposed take the tobacco leaf sensory evaluating smoking as basis, on the basis required clear and definite tobacco leaf aesthetic quality, by utilizing the Modern Instrument Analytical Technique means in conjunction with Chemical Measurement, obtain external appearance characteristic, physical characteristics to the tobacco leaf of research, physical and chemical index characteristic index of correlation parameter, utilize a series of mathematical statistics methods, finds factor of determination and difference factor to the tobacco leaf characteristic, explore the method for accurately feasible optimization quality index, for Continuous optimization and the correction of quality of tobacco system index provides approach.And, on the basis of clear and definite different places of production tobacco style and effect location, explore the impact of the main physical and chemical index of tobacco leaf on tobacco leaf aesthetic quality style, main physical and chemical index optimum range during the Optimizing Tobacco material quality requires, build the quality of tobacco evaluation system then.The quality that this method and evaluation system are estimated tobacco leaf has objectivity, systematicness and specific aim, and stronger operation instruction and practicality.

Concrete step is

(1) sample preparation

The sample obtained need to guarantee the diversity of sample physicochemical character " discreteness ", " wide in range property " and presentation quality, prepares sample and divides typical sample and outside two parts of checking sample of introducing.

The characteristics such as that the selected grade that requires a, typical sample contains is wide, purity is high, representative strong, this part sample detects the parameter of obtaining will be as basic assay data;

B, the outside checking sample of introducing, require originate wide, wide ranges, can represent that Cigarette Industrial Enterprise raw tobacco material buying level gets final product, and this part sample detects the parameter of obtaining will or revise data as checking;

C, when obtaining sample, correspondence is carried out the investigation of sampling spot ecology, kind and cultivation technique.

(2) pattern detection

A, by the sensory evaluating smoking, the tobacco leaf sample is carried out to feature description and industrial applicability analysis (odor type, fragrance matter, perfume quantity, comfortableness, assorted gas, concentration, pleasant impression, pungency etc.);

B, mensuration tobacco leaf sample presentation quality (color, colourity, identity, oil content, structure etc.);

C, measure sample physical characteristics (Dan Yechong, containing stalk rate, thickness, density etc.);

D, all kinds of chemical compositions of profile sample and the relevant index of quality (alkali four ratios, glucide, nitrogen-containing compound, alkaloid, organic acid, aldehydes matter, aroma substance etc.).

(3) former data are processed

The data resource of separate sources is carried out to the mathematics conversion by certain standard, make it particularly numerical data and be beneficial to statistical study.Extrapolate the indexs such as sugared alkali ratio, two sugar ratios, potassium chlorine ratio, two sugar differences.Simultaneously, rejecting abnormalities value.

(4) the sample essential characteristic is described and usability evaluation

From the layout in sample acquisition place, ecology and cultivation background, rate range and carry out the representativeness of typical tobacco sample, wide in range property and discreteness evaluation to detecting the aspects such as index normal distribution situation, each index essential characteristic description of quality of tobacco.The tobacco sample that only meets representativeness, wide in range property and discreteness just can carry out the quality of tobacco evaluation.

(5) application of evaluation index primary election and statistical method

Take the tobacco leaf sensory evaluating smoking as basis, utilize simple correlation, partial correlation and successive Regression, Double Selection successive Regression, path analysis, and the method such as artificial neural network carries out preliminary assessment and screening to each quality evaluation index, explore the main physical and chemical index that affects the quality of tobacco characteristic.

(6) determining of evaluation index and determining of optimum sample

Utilize the methods such as ordered sample Fisher classification or Hierarchical Clustering to carry out the division of cutting apart of sample and the quality of smokeing panel test to tobacco leaf sensory evaluating smoking quality.1) the affiliated sample of the segmentation of cutting apart is utilized to the test of hypothesis of multivariate analysis of variance single-factor and multiple ratio, further the screening and assessment index, determine relatively convenient, accurate and can reflect tobacco style and quality stability control index; 2) sample decomposition according to methods such as ordered sample Fisher classification or Hierarchical Clusterings, tobacco leaf sensory evaluating smoking quality carried out, in conjunction with central value ± SD(standard deviation), the confidence level that the numerical value interval of limiting is dropped in population sample is 68.2%, central value ± 1.96SD(standard deviation), be 95%, central value ± 2.57SD(standard deviation), be 99% requirement, determine optimum sample sensory evaluating smoking scope, then determine optimum sample.

(7) main physical and chemical index optimum range is just sentenced

According to determining optimum sample, tentatively obtain the optimum range of screened tobacco leaf evaluation index, their central value and franchise proposed.

(8) verification and debugging of optimum range

The definition sample number > 60% degree of conformity is restrictive condition, in conjunction with cigarette enterprise to the raw tobacco material quality requirements, introduce the external certificate sample, according to certain priority, the index screened is carried out to the degree of conformity checking, and then debugging and revision correlation technique parameter, determine the main physical and chemical index optimum range of tobacco leaf, with this, guarantee the dependable with function of optimum range.

(9) reckoning of ideotype growing way appearance and the measure that forms a complete production network

Obtain tobacco leaf sample in optimum range, carry out the essential characteristic description, find corresponding ecology, kind and cultivation technique, calculate cigarette strain ideotype growing way appearance and the measure that forms a complete production network.

(10) application of evaluation result

Take quality of tobacco as target, and ideotype growing way appearance and production measure are benchmark, carry out rational production distribution, select suitable flue-cured tobacco cultivars, formulate corresponding cultivation step, instruct tobacco commercial enterprise leaf tobacco production.Simultaneously, instruct buying and the utilization of Cigarette Industrial Enterprise raw tobacco material.And constantly sum up in application process, constantly improve evaluation method, revision evaluation index and optimum range, guarantee the advance of evaluation system.

Embodiment 2

1.1 test period, place

In 2007～2010 years, in Zunyi City, Guizhou Province, (Meitan, Fenggang), Bijie Prefecture (Bijie City, the west of Guizhou Province, generous, Jinsha) carried out this project, specifically in Table 1.

1.2 test material (kind)

K326, No. 3, Nan Jiang, cloud and mist 87, the relevant counties and cities of cloud and mist 97(company provide), specifically in Table 1.

1.3 the collection of sample

1.3.1 selection and plot experimental field settle the standard

2007:

Sampling plot, field settles the standard as follows:

(1) comprise high, medium and low three Cotton Varieties by Small Farming Households levels;

(2) soil fertility level comprises high, medium and low three levels;

(3) comprise field cigarette and native cigarette totally three factors.

In 4 characteristic tobacco production areas of Meitan County, target producing region, (encrinite, Gao Tai, Tian Cheng, wash 4 small towns of horse) selects sample representation cigarette ground, 12 blocks, different all kinds of cigarettes ground (3, every peasant household soil is determined in each production area, 1, field), and GPS location, by these cigarettes correspond to peasant household separately.When selecting cigarette ground, cigarette ground and tobacco grower are investigated simultaneously.Finally identify project and arranged encrinite, Gao Tai, Tian Cheng, wash horse 36 blocks of soil of 4 small towns 12Hu peasant households and 12 fields, be distributed in respectively 4, ,Bei He north and south, the Nan, southeast, Meitan direction.

2008:

Selected 3 peasant households on the basis of 07 annual selected peasant household, dwindled the roguing area, each peasant household selects 1 ground, and cigarette strain 300 strain left and right, in Table 2.Actual 80, the sample that obtains.

Emphasis sample collection (2007)

According to the layout of testing site and 48 collected plot cigarette samples, under Tang Yuanju researcher instructs, finally determine encrinite township period-luminosity build up a family fortune Gong Wen dragon man toft (medium production level, soil fertility inferior), day town and country Huang Chao Jun Jia behind the house 4 blocks of soil of head (inferior production level, soil fertility first-class) the tea La Shu of man of ,Xi Ma township Sun Wu unit (inferior production level, soil fertility first-class) ,Gao Tai town Li Kui literary composition man well bank (first-class production level, soil fertility medium) be that typical sample obtains plot.The production level of 3 grades in these 4 fixed point Han Gailiao upper, middle and lower, plot, 3 of upper, middle and lower soil fertility level, do not comprise field.The 333m2/ piece, adopt full receipts entirely, the sampling take " GB2635-92 flue-cured tobacco " be standard.Obtain altogether all good tobacco samples of degree of ripeness 282 Part, separate 36 grades.And can represent base leaf tobacco production level then.

Plot sample collection (2007)

After definite typical sample obtains plot, 44 plot of residue fixed point are adopted to full receipts equally entirely, sampling take " GB2635-92 flue-cured tobacco " be standard.Regularly sample, get grade X2F, C3F, C3L, C2F, C2L, B2F(C2F, C2L), obtain altogether sample 174 Part.

The collection of representative sample sample (2008)

In Meitan County, 7 tobacco leaf central stations select 14 purchase spots to be sampled (in Table 3), and sample comprises the main grade (in Table 4) of purchase, after the tobacco purchasing warehouse-in, carry out.Namely, in tobacco bale, select representational tobacco leaf 1.5kg from the required grade of sample.Planned sampling 294 Individual, the actual sample that obtains 289 Individual, contain 21 Individual grade.

Table 3 representative sample sampling tobacco leaf central station and purchase group name list

1.3.5 the large sample collection of threshing redrying plant

Sampling is carried out along two Redrying Factories of safe Ltd online in zunyi, guizhou Shen justice redrying Ltd and Guizhou Bijie.Beat leaf line paving leaf platform at Redrying Factory, according to overall number of samples, get at regular intervals one, cigarette sample, indicate tobacco leaf grade, source place and other relevant information.Sample comprises the whole grades of cigarette group at project tobacco leaf that adjust in the place of production, according to the big or small proportional sampling of each grade amount of adjusted tobacco leaf, guarantees authenticity and the representativeness of institute's sample thief as far as possible.Single cigarette sample is got 1.5 kilograms.

Year obtains tobacco sample altogether 138Part, 16 grades.

The yearly plan sampling 71Individual, 11 grades, actual sampling 60Individual, disappearance 11Individual.

Year sample collection

In Meitan, select representational 3 small towns as sampling spot, be respectively encrinite township (coptis root dam purchase group), south, Meitan County, town and country, sky, middle part (star connection purchase group), northern Xihe River township (two stone purchase group), sample on a small quantity in other small towns, Meitan flue-cured tobacco zone simultaneously.Fenggang County, the Jinsha County of Bijie, Qianxi County, Bijie City, four counties, Dafang County in Zun Yi are sampled.In the sampling of Meitan 21 grades (with table 4) altogether, the sampling grade in all the other each counties and cities is X2F, C3F, B3F Three Estate.Within 09 year, obtain altogether 190, sample.

The collection of this study sample be take main product Yan Xian Meitan, zunyi, guizhou city as main, is aided with Zunyi main product Yan Xian Fenggang and main product Yan Xian Bijie City, Bijie Prefecture, Jinsha, the west of Guizhou Province and the county such as generous amounts to 6 counties (city) approximately 26 Individual flue-cured tobacco main product small towns, take the mode of point (successively having operated 15 points) face combination, at beating and double roasting, also sampled online, by 2007,2008,2009 3 years, collects simultaneously 1133 Remaining part, almost contain the tobacco sample of congruent grade, the gained sample can represent the target producing region fully.Project is (1.3.2) emphasis sample and (1.3.4) representative sample, the verification sample that all the other samples be outside introducing.

Sample dissects

By Guizhou Province Tabacco Scientific Research Institute, 1133 collected tobacco samples are carried out the analyzing and testing of presentation quality, physical characteristics, main chemical compositions content, the quality of smokeing panel test and aroma component.

Evaluation on Appearance Quality

The presentation quality of tobacco leaf is the external feature of tobacco leaf, refer to the quality aspect that people's sense organ can judge, be the important evidence of tobacco leaf grading, formed by the size of tobacco leaf, homogeneity, integrality, foreign matter, residual wound, position, color (graininess, flexibility), identity (thickness, density), degree of ripeness, oil content, fragrance etc.These features are closely related with quality of tobacco, are the foundations that quality is divided.The at present domestic evaluation to Appearance Quality of Flue-cured Tobacco observing, the sense organ determination methods such as hand is touched, nose news, the qualitative description of general multiplex character property or concentrate on single proterties and other evaluation indexes between relation research, few people carry out thoroughgoing and painstaking, comprehensive and systematic quantitative test, to Appearance Quality of Flue-cured Tobacco comprehensive evaluation difficulty.Cai Xianjie etc. have selected 6 indexs such as color, degree of ripeness, blade construction, identity, oil content, colourity to be quantized and have inquired into the relation between these evaluation indexes; This studies cigarette district, base, above cigarette group Guizhou flue-cured tobacco is research object, as required, with reference to GB GB2635-92, has adopted the selected methods such as Cai Xianjie to carry out the evaluation of tobacco leaf presentation quality.Evaluation method is that elder generation's completely random from simple sample extracts 30, tobacco leaf, by sheet, makes the tobacco leaf Evaluation on Appearance Quality, calculates the identical appearance index number of sheets and accounts for the number percent of total number of sheets, and measure length and the width of these 30 leaves.

Wherein, sample appearance Quality Identification in 2007 work is completed by the rich old expert of Guizhou Province's quality of tobacco supervision measuring station presentation quality connoisseur's brocade celebrating, exterior quality is with text description, its factor comprises: position, color, degree of ripeness, identity, structure, oil content, colourity, grade, in Quality Identification, for describing more meticulously each quality factor feature, the intermediate value that basic composition is with each factor, if 3 levels, being greater than symbol for intermediate value "+" means, being less than symbol for intermediate value "-" means, as: the tobacco leaf for the tobacco leaf oil content at " having " class, can be subdivided into: " have ^-", " having ", " have ⁺", other exterior quality factors are by that analogy.For ease of data analysis, main quality factor color, oil content, colourity etc. are composed to a minute quantification, concrete quantization method is in Table 5, color from deep to shallow, oil content from more to less, colourity by dense in to weak tax score value, reduce gradually, score value only represents factor level, not the representation quality class.

Table 5 tobacco leaf presentation quality is composed a minute standard scale

Annotate: be greater than symbol for intermediate value "+" and mean, be less than symbol for intermediate value "-" and mean, between 3 levels, each level disparity is 0.5.

Physical index is estimated

The physical behavior of tobacco leaf comprises formalness and the physical property thereof of tobacco leaf.It is not only relevant to type, kind, grade and the quality of tobacco leaf, and with the establishment of the design of cigarette composition, tobacco leaf processing and storage technology, extremely close relationship is arranged, be the important indicator of reflection quality of tobacco and processing characteristics, directly affect product style, cost and other economic targets in quality of tobacco and cigarette manufacture process.In the purchase of China tobacco leaf grading and sales process, to the evaluation of tobacco leaf presentation quality, mainly with qualitative index, be foundation greatly.Because qualitative index can produce certain subjectivity randomness, this both had been unfavorable for instructing, and tobacco grower's production degree of ripeness is good, the tobacco leaf of Functionality, quality and appealing design, also be unfavorable for that tobacco enterprise is correctly used tobacco leaf in cigarette composition design, tobacco leaf processing and storage process, improves utilization ratio and the economic worth of tobacco leaf.For making tobacco leaf presentation quality and industrial applicability estimate more science, reasonable, system and comprehensive, cigarette industry adds the index of quantifiable physical behavior in to the tobacco leaf Evaluation on Appearance Quality.Originally research and propose long with leaf, leaf is wide, degree of hacking, Dan Yechong, surface density (unit leaf area weight), reflect the physical behavior index of quality of tobacco and industrial applicability containing stalk rate etc., and above cigarette group Guizhou flue-cured tobacco leaves at main areas sample is material, the tobacco leaf physical behavior provincial characteristics in main product cigarette district, base, cigarette group Guizhou in in-depth analysis, to its carry out relevant evaluation and with other index related analyses.

1. leaf is long, leaf is wide (%): length of blade is measured piecewise, and not enough 1cm presses 1cm and calculates, the length that the average of length of blade is this sample; Width of blade is measured piecewise, and not enough 0.5cm presses 0.5cm and calculates, the width that the average of width of blade is this sample.

2. single leaf weighs (g), leaf density (mgcm-3): single leaf heavily refers to the weight of a slice leaf.Randomly draw the tobacco leaf that 10 water percentage are 15% left and right, every tobacco leaf is appointed and is got one and half leaves, in the blade tip of half leaf, leaf and leaf base is equidistant gets 5 points, make a call to 5 diameters (D) for the circular shaped patches of 15mm with circular card punch, 50 circular shaped patches are put into to the moisture box, dry 2h under 100 ℃ of conditions, weigh after cooling 30min.Leaf density refers to the tobacco leaf weight of unit volume.

3. containing stalk rate (%): randomly draw 20 tobacco leaves, equilibrium moisture content is to 16.5% scholar's 0.5% tobacco leaf, take out stalk, then with 1/100 balance, divide the weight of another name smoked sheet and offal, press formula and calculate containing the stalk rate: containing stalk rate=(offal weight/tobacco leaf weight) * 100.

Tobacco leaf chemical composition is analyzed

Tobacco leaf chemical composition refer to organism and inorganics in tobacco leaf the content height and between proportionate relationship, be the important evidence of weighing quality of tobacco.While due to the flue gas of people suction, being the leaf chemical ingredient burning, through distillation, destructive distillation, pyrolysis, produce, tobacco leaf chemical composition affect flue gas characteristic, thereby tobacco leaf chemical composition can be used as the index of evaluation quality of tobacco.The chemical composition of tobacco leaf and tobacco type, cultivation, modulation and processing have close relationship.At present, the tobacco leaf that oneself identifies and smoke components have 5289 kinds, along with the progress of analytical technology, also have some micro constitutents and labile element constantly to be identified out.Tobacco leaf chemical composition is very complicated, and which kind of chemical index is incapability use, and all can not reflect comprehensively or express quality of tobacco, but tobacco leaf intrinsic chemical composition remains one of important indicator of estimating Flue-cured tobacco Quality.At present, less for the report of upper cigarette group Guizhou base Chemical Compositions of Flue-cured Tobacco research.This research as required, will be to seven index total reducing sugars of routine, reducing sugar, nicotine, total nitrogen, potassium, chlorine, the protein of tobacco leaf chemical composition, and tobacco leaf harmony index sugar/alkali, two sugar ratio, nitrogen/alkali, potassium/chlorine and two sugar poor etc. carry out relevant evaluation and with other index related analyses, and between the tobacco leaf main physical and chemical index Suitable Area of acquisition for upper cigarette group cigarette brand.

Chemical composition analysis detects according to being respectively: YC/T159-2002 (total reducing sugar), YC/T159-2002 (reducing sugar), YC/T160-2002 (total alkaloid), YC/T161-2002 (total nitrogen), YC/T173-2003 (potassium), YC/T162-2002 (chlorine), protein determination adopts IKI colorimetric continuous flow method.Each detects data and all is converted into percent.

1.4.4 the tobacco leaf organoleptic quality is estimated

The flue-cured tobacco quality of smokeing panel test, claim again the inherence quality of smokeing panel test, refer to that tobacco leaf passes through the feature of the burning flue gas that produces, mainly rely on the personnel's suction of smokeing panel test to be marked to identify to its main evaluation index, though with certain subjectivity, but still be the main reference frame that cigarette enterprise is used certain tobacco leaf and formula Design.In the factors of weighing the flue gas quality, fragrance and suction flavor are topmost.Fragrance comprises odor type, fragrance matter, perfume quantity and assorted gas; Inhale flavor and comprise strength (physiological strength), pungency, concentration and pleasant impression etc.

The evaluation of tobacco leaf organoleptic quality is carried out according to a conventional method, and the content of smokeing panel test mainly comprises fragrance matter, perfume quantity, jealous, assorted gas, pungency, strength, flammability and grey etc.By collected specimens, by single regional chopping, perfuming is reinforced, with the same obbbin, through machinery, rolls into all consistent cigarettes of length, thickness, degree of tightness.More than equilibrium moisture 24h, then adopt the inherent quality of the method judge single-tobacco-typed cigarette sample of smokeing panel test of whole circulation under the condition of 65% relative humidity.Regulation with reference to ((YC/T138-1998 tobacco and tobacco product " and " HTTS/QPM3-17)) standard, single-tobacco-typed cigarette smoke panel test quality index and standards of grading have been set up, the expert that smokes panel test by the Guizhou Province Tabacco Scientific Research Institute Ping Yan council, carry out the sensory evaluating smoking according to standard.According to 8 indexs such as fragrance matter, perfume quantity, jealous, assorted gas, pungency, strength, flammability and grey, the sample of smokeing panel test is given a mark respectively, all experts of then usining smoke panel test the mean value of scoring as final smoking result.

1.5 constituent analysis and statistical method

The analytical approach of Chemical Components of Tobacco Leaves total alkaloid, soluble sugar, reducing sugar, total nitrogen, potassium, chlorinity adopts the near infrared spectroscopic method of Shanghai Tobacco's unified Modeling.Sugar alkali ratio, two sugar ratios, nitrogen base ratio and potassium chlorine are the reckoning value than ratio.

Use the methods such as descriptive statistics, simple correlation, partial correlation, successive Regression, principal component analysis (PCA) and artificial neural network to carry out statistical study to data.Take the tobacco leaf smoking result as basis, ocular estimate index, chemical composition index and new every evaluation index of introducing to tobacco leaf are screened, clear and definite they on the size that affects of tobacco leaf feature, divide between the Suitable Area of the chemical index that significant contribution is arranged, their central value and franchise proposed.With reference to Shanghai cigarette industry raw tobacco material quality evaluation system (revision for the third time in 2005) evaluation Chemical Components of Tobacco Leaves content and harmony.

2 results and analysis

2.1 the sample cluster quality index is estimated

2.1.1 the evaluation of sample group quality index is mainly to estimate it: representative, wide in range property and discreteness.

From obtained sample, will basically in the ratio of 1:1:1, be carried out by the sample cluster upper, middle and lower section that be applied to analyze, and obtain enough large sample.

2.1.2 sampling spot is arranged, ecological and cultivation context analyzer evaluation

From the sampling spot layout of sample, ecological, cultivation background, the sampling spot of 2007 is distributed in four main product cigarette small towns of 4 directions in ,Bei He north and south, the Nan, southeast, Meitan, within 2008, sampling spot is contained 14 purchase groups of the whole tobacco leaf central stations of Meitan tobacco branch office (7), the whole ecological condition that has comprised Meitan, the production technology peasant household of different habitant level and different levels, and according to local sound tobacco production technology, cultivated fully, the gained sample can well represent target producing region cured tobacco production level and quality of tobacco.

 

By sample cluster being made to descriptive statistical analysis, can draw: in sample cluster, all of emphasis plot mean is 2.4959, standard deviation is 0.88851, coefficient of kurtosis is-0.032, the standard of coefficient of kurtosis is mistaken for 0.289, and the coefficient of skewness is 0.427, and the standard of the coefficient of skewness is mistaken for 0.145, minimum value is 0.76, maximal value is 5.48, and effectively number of cases is 282, without missing values.In sample cluster, all mean of bulk production is 2.7668, standard deviation is 1.08363, coefficient of kurtosis is 0.115, the standard of coefficient of kurtosis is mistaken for 0.410, and the coefficient of skewness is 0.653, and the standard of the coefficient of skewness is mistaken for 0.206, minimum value is 0.65, maximal value is 5.74, and effectively number of cases is 138, without missing values.The representative sample mean that in sample cluster, 08 year obtains is 2.1491, standard deviation is 0.99411, coefficient of kurtosis is 0.422, the standard of coefficient of kurtosis is mistaken for 0.288, and the coefficient of skewness is 0.972, and the standard of the coefficient of skewness is mistaken for 0.144, minimum value is 0.58, maximal value is 5.35, and effectively number of cases is 285, without missing values.

By the normal distribution to each position sample of sample cluster (this example province), Meitan, Guizhou, the target producing region sample obtained meets normal distribution.Come from the congruence level sampling on face at 08 year sample cluster that obtains itself, here, we even can think that 08 year sample directly has its representativeness.

Rule of thumb and general knowledge judgement, a good random sampling sample group and better meet the wide in range property of quality index and sample group that discreteness requires should meet and approach normal distribution.By normal distribution and to the scatter diagram of color sample, Dan Yechong, nicotine, sugar/alkali etc., the sample cluster obtained, leading indicator has possessed the characteristics of " wide in range property " and " discreteness ".

From above correlation graph analysis, in conjunction with sample hierarchical organization and sampling spot layout, ecology etc., we think that sample group's the quality index of preparation has represented certain bulk production level, show that sample cluster has good representativeness, wide in range property and discreteness, gained sample group energy well represents target producing region tobacco leaf.

Sample group quality index feature

2.2.1 sample cluster presentation quality

As can be seen from Table 10, in 160 bottom leaf sample clusters, degree of ripeness (being 95 samples) mean scores is 8.75 minutes, and the coefficient of variation is 7.55%; Color average is 5.90 minutes, and the coefficient of variation is 26.77%; Oil content mean value is 5.33 minutes, shows that oil content is slightly having class, and the coefficient of variation is 13.64%; Identity mean value is 7.45 minutes, and indicate identification is at slightly thin class, and the coefficient of variation is 14.26%; Structure mean value is 8.95, shows that structure is at loose class, and the coefficient of variation is 3.85%; Colourity mean value is 4.45, shows that colourity is at weak extremely middle class, and the coefficient of variation is 19.81%.

As can be seen from Table 10, in 176 middle leaf sample clusters, degree of ripeness (being 95 samples) mean scores is 8.96 minutes, and the coefficient of variation is 1.35%; Color average is 6.57 minutes, and the coefficient of variation is 17.53%; Oil content mean value is 6.46 minutes, shows that oil content is slightly having to class is arranged, and the coefficient of variation is 8.82%; Identity mean value is 8.40 minutes, and indicate identification slightly is being as thin as medium class, and the coefficient of variation is 6.35%; Structure mean value is 8.87, shows that structure is at loose class, and the coefficient of variation is 4.65%; Colourity mean value is 5.79, shows that colourity is at middle class, and the coefficient of variation is 12.99%.

As can be seen from Table 10, in 184 upper leaf sample clusters, degree of ripeness (being 90 samples) mean scores is 8.47 minutes, and the coefficient of variation is 7.90%; Color average is 6.02 minutes, and the coefficient of variation is 31.10%, and the coefficient of variation is larger; Oil content mean value is 6.16 minutes, shows that oil content is slightly having class, and the coefficient of variation is 13.26%; Identity mean value is 7.02 minutes, and indicate identification is at slightly thick class, and the coefficient of variation is 13.39%; Structure mean value is 7.14, shows that structure is at slightly close class, and the coefficient of variation is 10.38%; Colourity mean value is 5.09, shows that colourity is at middle class, and the coefficient of variation is 20.15%.

2.2.2 sample cluster physical arrangement

As can be seen from Table 11, in 160 bottom leaf sample clusters, leaf length between 35.41-67.66cm, mean value 50.36cm, the coefficient of variation is 12.82%; Leaf is wide between 12-24.39cm, and mean value is 17.55cm, and the coefficient of variation is 13.74%; Single leaf focuses on 2.14-11.50g between, average out to 5.09g, the coefficient of variation is 34.88%; Contain the stalk rate between 23.88-42.86%, average out to 31.67%, the coefficient of variation is 11%; Surface density is between 26.65-76.08g/m2, and mean value is 44.49g/m2, and the coefficient of variation is 17.47%.

As can be seen from Table 11, in 176 middle leaf sample clusters, leaf length between 47.58-74.24cm, mean value 60.26cm, the coefficient of variation is 8.33%; Leaf is wide between 15.55-27.72cm, and mean value is 19.33cm, and the coefficient of variation is 12.09%; Single leaf focuses between 4.26-15.92g, average out to 8.48g, and the coefficient of variation is 25.94%; Contain the stalk rate between 26.85-40.19%, average out to 31.98%, the coefficient of variation is 7.62%; Surface density is between 35.01-91.62g/m2, and mean value is 59.54g/m2, and the coefficient of variation is 18%.

As can be seen from Table 11, in 184 upper leaf sample clusters, leaf length is between 43.7-69.5cm, and mean value is 55.82cm, and the coefficient of variation is 10.07%; Leaf is wide between 10.87-20.52cm, and mean value is 15.42cm, and the coefficient of variation is 11.68%; Single leaf focuses between 3.99-16.7g, average out to 8.11g, and the coefficient of variation is 26.53%; Contain the stalk rate between 21.04-37.39%, average out to 28.27%, the coefficient of variation is 9.91%; Surface density is between 51.18-108.74g/m2, and mean value is 81.08g/m2, and the coefficient of variation is 15.42%.

2.2.3 sample cluster chemical composition

Table 12 Chemical Components of Tobacco Leaves is described (Descriptive Statistics)

From table 12, in 160 bottom leaf samples, bottom leaf nicotine between 0.73-3.6%, average out to 1.63%, the coefficient of variation is 29.63; Reducing sugar is between 0.56-30.59%, and mean value is 17.13%, and the coefficient of variation is 38.23%; Total nitrogen is between 1.33-3.08%, and mean value is 2.04%, and the coefficient of variation is 15.17%; Potassium content is between 1.18-3.62%, and mean value is 2.10%, and the coefficient of variation is 21.04%; Protein content is between 5.45-11.3%, and mean value is 7.71%, and the coefficient of variation is 12.7%; Sugar/alkali content is between 0.77-35.65, and mean value is 11.53, and the coefficient of variation is 52.94%; Two sugared ratio contents are between 0.12-1.00, and mean value is 0.87, and the coefficient of variation is 12.62%; Nitrogen/alkali content is between 0.56-3.51, and mean value is 1.34, and the coefficient of variation is 27.97%; Potassium/chlorinity is between 2.35-13.91, and mean value is 5.72, and the coefficient of variation is 35.52%.

From table 12, in 176 middle leaf samples, nicotine between 0.84-3.78%, average out to 2.15%, the coefficient of variation is 29.71%; Reducing sugar is between 8.36-31.90%, and mean value is 24.21%, and the coefficient of variation is 16.41%; Total nitrogen is between 1.28-2.98%, and mean value is 1.95%, and the coefficient of variation is 14.4%; Potassium content is between 1.08-3.02%, and mean value is 1.84%, and the coefficient of variation is 19.67%; Protein content is between 5.49-10.55%, and mean value is 7.07%, and the coefficient of variation is 11.51%; Sugar/alkali content is between 2.80-34.20, and mean value is 12.73, and the coefficient of variation is 44.62%; Two sugared ratio contents are between 0.76-1, and mean value is 0.92, and the coefficient of variation is 3.52%; Nitrogen/alkali content is between 0.55-2.2, and mean value is 0.98, and the coefficient of variation is 30.56%; Potassium/chlorinity is between 1.08-12.8, and mean value is 5.71, and the coefficient of variation is 30.9%.

From table 12, in 184 upper leaf samples, nicotine between 0.58-5.35%, average out to 2.93%, the coefficient of variation is 35.73%; Reducing sugar is between 0.53-31.39%, and mean value is 21.31%, and the coefficient of variation is 25.01%; Total nitrogen is between 1.27-3.75%, and mean value is 2.22%, and the coefficient of variation is 24.62%; Potassium is between 0.35-2.55%, and mean value is 1.41%, and the coefficient of variation is 28.26%; Protein is between 5.2-12.05%, and mean value is 7.56%, and the coefficient of variation is 19.5%; Sugar/alkali is between 0.12-48.24, and mean value is 9.76, and the coefficient of variation is 84.53%; Two sugar are than between 0.81-1.00, and mean value is 0.94, and the coefficient of variation is 3.92%; Nitrogen/alkali is between 0.53-3.21, and mean value is 0.85, and the coefficient of variation is 44.77%; Potassium/chlorine is between 0.61-8.97, and mean value is 3.54, and the coefficient of variation is 36.09%.

2.2.4 sample cluster sensory evaluating smoking

From table 13, in 160 bottom leaf samples, aesthetic quality's characteristic aspect fragrance matter is between 6.6-8, and mean value is 7.55, and the coefficient of variation is 3.5%; Perfume quantity is between 6.8-8, and mean value is 7.59, and the coefficient of variation is 2.23%; Jealous between 7-9, mean value is 8.26, and the coefficient of variation is 4.79%; Assorted gas is between 6.2-8, and mean value is 7.43, and the coefficient of variation is 4.11%; Pungency is between 6.8-8, and mean value is 7.53, and the coefficient of variation is 2.5%; Always assign to see, between 34.3-40.7, average out to 38.35, the coefficient of variation is 3.23%.

From table 13, in 176 middle leaf samples, fragrance matter is between 6.60-8.40, and mean value is 7.86, and the coefficient of variation is 3.87%; Perfume quantity is between 7.10-8.40, and mean value is 7.98, and the coefficient of variation is 2.86%; Jealous between 7.10-9.5, mean value is 8.49, and the coefficient of variation is 4.79%; Assorted gas is between 6.40-8.20, and mean value is 7.60, and the coefficient of variation is 3.76%; Pungency is between 6.80-8.1, and mean value is 7.62, and the coefficient of variation is 2.83%; Always assign to see, between 34.20-42.30, average out to 39.55, the coefficient of variation is 3.39%.

From table 13, in 184 upper leaf samples, fragrance matter is between 6.1-8.30, and mean value is 7.54, and the coefficient of variation is 5.03%; Perfume quantity is between 6.8-8.40, and mean value is 7.88, and the coefficient of variation is 4.31%; Jealous between 6.5-8.90, mean value is 8.08, and the coefficient of variation is 5.78%; Assorted gas is between 5.50-7.90, and mean value is 7.30, and the coefficient of variation is 5.01%; Pungency is between 5.80-7.80, and mean value is 7.32, and the coefficient of variation is 3.87%; Always assign to see, between 30.9-41.30, average out to 38.14, the coefficient of variation is 4.57%.

By the analysis to sample cluster, in general, Meitan tobacco leaf degree of ripeness, color and colourity upper, middle and lower section coefficient of variation aspect presentation quality is all larger, on homogeneity the performance poor, all the other indices in general homogeneity are better.The single leaf in physical arrangement aspect is heavy and the surface density upper, middle and lower section coefficient of variation is all larger, on homogeneity, shows poorly, and all the other indices in general homogeneity are better.The overall homogeneity in chemical composition aspect is poor, indices wider distribution in sample cluster, but from each position nicotine mean value, under, in, top nicotine distribution table is now very good, is followed successively by 1.63%, 2.15%, 2.93%.The sensory evaluating smoking aspect, under, in, top all shows more unanimously, the coefficient of variation all<6, can say that it is very desirable that the Meitan tobacco leaf shows aspect smokeing panel test.

Correlation analysis between the quality of tobacco constituent element

In correlation analysis, usually utilize related coefficient to analyze or measure the linear dependence degree between variable.Yet simple correlation coefficient is subject to the impact of other factors, the nonessential contact often of reflection.Partial correlation coefficient is under the condition of being controlled in the impact on its dependent variable, weighs the index of the linear dependence degree between certain two variable in a plurality of variablees, can reflect more realistically the degree of correlation between variable.This research adopts partial correlation coefficient and principal component analysis (PCA) to describe between correlativity between the quality of tobacco constituent element and each index thereof the contribution rate to the quality of tobacco constituent element, according to the accumulation contribution rate > 85% principle, extract relevant major component.

2.3.1 the analysis between presentation quality

Data from table 14, partial correlation performance lower tobacco leaf is negative correlation except structure and identity, colourity, between all the other each indexs, shows as positive correlation, color and identity, structure, oil content and identity, colourity, the mutual conspicuousness of identity and colourity is obvious.Analysis result shows, the quantity of information that front 3 principal components form is gross information content 85.78%, almost reflected full detail.In these 3 principal components: the 1st principal component is except color and structure, and all the other each composition weight coefficients are all comparatively approaching, the 1st principal component composition contribution rate 50.08%; The 2nd principal component representative structure, weight coefficient is 0.8735, contribution rate is 22.00%; The 3rd principal component representative color, weight coefficient is 0.7288.

Data from table 15, partial correlation coefficient shows that middle part tobacco leaf oil content is negative correlation with color, structure, and oil content and structure be extremely remarkable negative correlation, between all the other each indexs, shows as positive correlation, wherein, identity and oil content, structure related coefficient are larger.Analysis result shows, the quantity of information that front 4 principal components form is gross information content 91.51%, almost reflected full detail.In these 4 principal components: the 1st principal component identity and colourity weight coefficient are comparatively approaching, composition contribution rate 39.94%; The 2nd principal component represents oil content and structure, and the composition contribution rate is 20.72%; The 3rd principal component representative color, weight coefficient is 0.8439, the composition contribution rate is 17.60%.

Data from table 16, partial correlation shows that upper tobacco leaf identity and color, colourity are outside negative correlation, between all the other each indexs, all shows as positive correlation, wherein identity and oil content, colourity and color, oil content, structure related coefficient are larger.Analysis result shows, the quantity of information that front 3 principal components form is gross information content 82.96%, almost reflected full detail.In these 3 principal components: the 1st principal component is except identity, and all the other each composition weight coefficients are all comparatively approaching, its composition contribution rate 49.14%; The 2nd principal component representative capacity, weight coefficient is 0.8313, its composition contribution rate is 20.93%; The 3rd principal component represents oil content and structure.

2.4.2 the analysis between physical characteristics

Between the tobacco leaf physical characteristics, data from table 17, lower tobacco leaf is containing the stalk rate and leaf is wide, Dan Yechong, surface density, and surface density and leaf are long, the wide negative correlation that is of leaf, all show as utmost point marked positive correlation relation between all the other each indexs.Analysis result shows, the quantity of information that front 3 principal components form is gross information content 96.61%, almost reflected full detail.In these 3 principal components: the 1st the principal component leaf is long and the heavy weight coefficient of single leaf is comparatively approaching, is respectively 0.5631 and 0.5700, its composition contribution rate 57.98%; The 2nd principal component representative is containing the stalk rate, and weight coefficient is 0.7283, composition contribution rate 25.67%.

Data from table 18, middle part tobacco leaf leaf is wide with leaf length, containing the stalk rate, surface density and leaf are long, leaf is wide, containing the stalk rate, be negative correlation, all show as positive correlation between all the other each indexs, long, the surface density wide with leaf of disleaf, contain stalk rate and Dan Yechong without outside remarkable relation, be extremely significantly relation between all the other indexs.Analysis result shows, the quantity of information that front 3 principal components form is gross information content 95.49%, almost reflected full detail.In these 3 principal components: the 1st the principal component leaf is long and the heavy weight coefficient of single leaf is comparatively approaching, is respectively 0.5333 and 0.5975, its composition contribution rate 52.70%; The 2nd principal component representative is containing the stalk rate, and weight coefficient is 0.6899, composition contribution rate 29.68%; The 3rd principal component represents Ye Kuan, and weight coefficient is 0.7215.

Data from table 19, upper tobacco leaf is wide containing stalk rate and leaf, surface density and leaf are long, leaf is wide, containing the stalk rate, be negative correlativing relation, show as positive correlation between all the other each indexs, wherein, Dan Yechong and leaf are long, leaf is wide, and long containing stalk rate and leaf, page density and leaf are long, leaf is wide, Dan Yechong, contain the stalk rate larger correlativity is arranged.Analysis result shows, the quantity of information that front 3 principal components form is gross information content 95.24%, almost reflected full detail.In these 3 principal components: the 1st the principal component leaf is long, leaf is wide and the heavy weight coefficient of single leaf is comparatively approaching, its composition contribution rate 54.22%; The 2nd principal component is comparatively approaching containing stalk rate and surface density weight coefficient, is respectively 0.6741 and 0.6982, composition contribution rate 34.85%.

2.4.3 the analysis between Chemical Components of Tobacco Leaves

The routine chemical components aspect, from table 20, bottom nicotine and total nitrogen are utmost point marked positive correlation, with protein, are extremely significantly negative correlativing relation, with total reducing sugar, potassium, chlorine, are negative correlation, not remarkable, with reducing sugar, are negative correlation.Total reducing sugar and reducing sugar are utmost point marked positive correlation, with protein, are remarkable negative correlation, and all the other are index related not remarkable.Total nitrogen and potassium are extremely significantly negative correlation, with the remarkable negative correlation of chlorine, be utmost point marked positive correlation with protein, and related coefficient are larger.Protein and potassium chlorine are utmost point marked positive correlation relation.Analysis result shows, the quantity of information that front 3 principal components form is gross information content 87.34%.In these 3 principal components: the 1st principal component is except nicotine, potassium and chlorine, and between all the other each indexs, weight coefficient is comparatively approaching, its composition contribution rate 52.74%; The 2nd principal component nicotine and potassium approach, and weight coefficient is respectively 0.6641,0.6160, composition contribution rate 21.97%; The 3rd principal component represents chlorine, and weight coefficient is 0.9227.

From table 21, middle part nicotine and total nitrogen are utmost point marked positive correlation, with protein, are extremely significantly negative correlation.Total reducing sugar and reducing sugar are utmost point marked positive correlation, and related coefficient reaches 0.9510, with chlorine, are extremely significantly negative correlativing relation.Reducing sugar is significantly relevant to potassium, chlorine.Total nitrogen and potassium significant correlation, and protein is utmost point clear-cut correlation.Potassium and chlorine are remarkable negative correlation.Analysis result shows, the quantity of information that front 3 principal components form is gross information content 82.64%.In these 3 principal components: between the 1st principal component two sugar and each index of total nitrogen, weight coefficient is comparatively approaching, its composition contribution rate 52.21%; The 2nd principal component nicotine and potassium weight coefficient approach, and are respectively 0.6987 and 0.6543, and the composition contribution rate is 15.61%; The 3rd principal component represents chlorine, and weight coefficient is 0.9028.

From table 22, top nicotine and total reducing sugar, total nitrogen are utmost point marked positive correlation, with reducing sugar and protein, are extremely significantly negative correlation.Be utmost point marked positive correlation between two sugar, total reducing sugar and total nitrogen, chlorine are extremely significantly negative correlativing relation.Reducing sugar and total nitrogen, chlorine are remarkable relation.Protein and total nitrogen, potassium are utmost point marked positive correlation relation.The analysis result demonstration, the quantity of information that front 3 principal components form reaches 92.28% of gross information content.In these 3 principal components: between the 1st principal component two sugar and each index of total nitrogen, weight coefficient is comparatively approaching, its composition contribution rate 69.69%; The 2nd principal component represents potassium, and weight coefficient is 0.8965, and the composition contribution rate is 13.84%; The 3rd principal component represents chlorine, and weight coefficient is 0.9038.

The analysis between quality 2.4.4 tobacco leaf is smoked panel test

The quality of smokeing panel test aspect, from table 23, between each index of bottom, perfume quantity and pungency are outside negative correlation, between all the other indexs, are positive correlation.The analysis result demonstration, the quantity of information that front 1 principal component forms reaches 86.01% of gross information content, has comprised all indexs, and weight coefficient approaches.

From table 24, be negative correlativing relation equal being proportionate property between all the other each indexs between each index of middle part except perfume quantity and assorted gas, pungency.The analysis result demonstration, the quantity of information that front 1 principal component forms reaches 85.83% of gross information content, and except perfume quantity and pungency, between all the other each indexs, weight coefficient is comparatively approaching, and all the other composition contribution rates are for less.

From table 25, fragrance matter and pungency between each index of top, perfume quantity and assorted gas are negative correlation, being proportionate property between all the other indexs.The analysis result demonstration, the quantity of information that front 1 principal component forms reaches 89.75% of gross information content, and this 1 principal component has contained all indexs, and between each index, weight coefficient is comparatively approaching.

Upper, middle and lower section all shows as fragrance matter and perfume quantity, assorted gas, jealous and assorted gas, pungency, and assorted gas and the pungency correlativity larger, be extremely remarkable relation.

2.4.5 Chemical Components of Tobacco Leaves and aesthetic appearance correlation among traits

Pass through stepwise regression analysis, as can be seen from Table 26, nicotine and bottom identity, middle part color, leaf length, Dan Yechong and surface density, upper leaf is long, leaf is wide, Dan Yechong and surface density are utmost point marked positive correlation, and, top identity wide with substructure, middle leaf is extremely significantly negative correlation.Total reducing sugar, reducing sugar and under, central structure, containing the stalk rate, the middle part color, superstructure, surface density are negative correlation, with middle part, containing the stalk rate, upper leaf is long, leaf is wide, Dan Yechong, containing the stalk rate, be extremely significantly negative correlation, and is utmost point marked positive correlation between all the other each indexs.Total nitrogen and middle part color and containing stalk rate, Dan Yechong, upper leaf is long, leaf is wide, Dan Yechong and be utmost point marked positive correlation containing the stalk rate, and with bottom color, oil content, identity, colourity, leaf is wide and single leaf weight, top color, identity are extremely significantly negative correlation.Potassium and central structure, containing the stalk rate, upper leaf is wide and be utmost point marked positive correlation containing the stalk rate, with bottom colourity and surface density, middle part oil content, Dan Yechong and surface density, top colourity is extremely significantly negative correlation.Chlorine and the bottom leaf is long, leaf is wide, Dan Yechong and containing the stalk rate, top identity and colourity are extremely significantly negative correlation.Protein and bottom color, oil content, identity, colourity, leaf is long, leaf is wide, Dan Yechong and surface density, and top color, colourity are extremely significantly negative correlation, and with middle part, containing the stalk rate, upper leaf is long, leaf is wide, Dan Yechong and surface density are utmost point marked positive correlation.From the indices correlativity, color, oil content, identity, colourity, and Dan Yechong, higher containing the correlativity of stalk rate, surface density and chemical composition.

2.4.6 the correlativity between each physical and chemical index of tobacco leaf and aesthetic quality

As can be seen from Table 27, according to stepwise regression analysis, between color (fragrance matter and perfume quantity) oil content, identity, colourity, surface density, total reducing sugar, reducing sugar, sugared alkali ratio, two sugar ratios and each index of aesthetic quality, be utmost point marked positive correlation relation.Be extremely significantly negative correlativing relation between total nitrogen, protein, two sugar poor (part) and each index of aesthetic quality.Leaf is grown, is contained between stalk rate (part is extremely remarkable), nicotine, potassium (removing and perfume quantity), nitrogen base ratio, potassium/chlorine and each index of aesthetic quality and is negative correlativing relation.Oil content, identity, colourity, surface density, total reducing sugar, reducing sugar, total nitrogen, protein, sugar/alkali, two sugar ratios, two sugar differences are extremely significantly relation with the total points of smokeing panel test.

From table 28, we can find out, according to stepwise regression analysis, middle part tobacco leaf oil content, identity (part significantly), leaf is long, leaf is wide, Dan Yechong, surface density, total reducing sugar, reducing sugar, sugar/alkali, two sugar be than with each index of aesthetic quality, being utmost point marked positive correlation, be extremely significantly negative correlation containing stalk rate, nicotine, total nitrogen, potassium (part significantly), protein (part significantly) with each index of aesthetic quality, each index of chlorine and aesthetic quality is negative correlation.Colourity and potassium/chlorine with each index of aesthetic quality, be proportionate.Oil content, identity, leaf is long, leaf is wide, Dan Yechong, containing stalk rate, surface density, nicotine, total reducing sugar, reducing sugar, total nitrogen, potassium, protein, sugar/alkali, two sugar than with the total points of smokeing panel test, being extremely significantly relation.

As can be seen from Table 29, according to stepwise regression analysis, between tobacco leaf color (fragrance matter, perfume quantity), oil content, identity, total reducing sugar, reducing sugar, sugar/alkali, potassium/chlorine, two sugar differences and aesthetic quality's indices, there is utmost point marked positive correlation relation.There is extremely significantly negative correlation between each index of nicotine, total nitrogen, chlorine, protein and aesthetic quality.Leaf length, Dan Yechong, containing stalk rate, surface density, two sugar than and each index of aesthetic quality between be negative correlation.Structure, colourity, leaf are wide, be proportionate between potassium (and perfume quantity is negative correlation) and each index of aesthetic quality.Color, oil content, identity, nicotine, total reducing sugar, reducing sugar, total nitrogen, potassium, protein, sugar/alkali, potassium/chlorine, two sugar differences are extremely significantly relation with the total points of smokeing panel test.

Correlationship analysis between comprehensive each physical and chemical index of above tobacco leaf and aesthetic quality's index, tobacco leaf color, oil content, identity, leaf are wide, Dan Yechong, surface density, nicotine, total reducing sugar, reducing sugar, total nitrogen, potassium, protein, sugared alkali ratio, potassium chlorine ratio, two sugar differ from and aesthetic quality's correlativitys are higher.

2.5 the division of the suitableeest quality of smokeing panel test of tobacco leaf

The division of the quality of smokeing panel test 2.5.1 the suitableeest

Adopt ordered sample Fisher sorting technique, the tobacco leaf sample the suitableeest quality of smokeing panel test in target producing region has been carried out to following division.So-called ordered sample Fisher sorting technique refers to the sort research of carrying out under the prerequisite of not upsetting the sample data sequence order.

From table 30, adopted ordered sample Fisher sorting technique to be divided the bottom leaf quality of smokeing panel test, we are tentative, and bottom Ye Zuishi smokes panel test quality between 38.7-40.7, and it accounts for bottom population sample group 48.13%.

From table 31, adopted ordered sample Fisher sorting technique to be divided the middle leaf quality of smokeing panel test, our tentative the suitableeest quality of smokeing panel test of middle leaf is between 39.9-42.3, and it accounts for middle part population sample group 46.02%.

From table 32, adopted ordered sample Fisher sorting technique to be divided the upper leaf quality of smokeing panel test, our tentative the suitableeest quality of smokeing panel test of upper leaf is between 37.9-41.3, and it accounts for top population sample group 67.39%.

Equally, we can take Hierarchical Clustering to be divided, and its division result is consistent with the optimal parting of ordered sample method;

2.5.2 the main physical and chemical index diversity ratio of typical sample of the different score values of smokeing panel test

By adopting ordered sample Fisher sorting technique, smoke panel test each physical and chemical index of sample of score value (section) of the difference of gained is carried out to multivariate analysis of variance and single-factor test of hypothesis:

From table 33, oil content, colourity, total reducing sugar, reducing sugar, total nitrogen, sugar/alkali, two sugar ratios and the poor utmost point significant differences that exist of two sugar between the different score values in bottom, identity, leaf be wide, containing stalk rate, surface density, protein, have significant difference, and other indicator differences are not remarkable;

From table 34, oil content between the different score values in middle part, leaf is long, leaf is wide, Dan Yechong, containing stalk rate, surface density, nicotine, total reducing sugar, reducing sugar, total nitrogen, potassium, sugar/alkali and two sugar than there being utmost point significant difference, there is significant difference in identity, and other indicator differences are not remarkable;

From table 35, between the different score values in top, there are utmost point significant difference in color, oil content, identity, colourity, nicotine, total reducing sugar, reducing sugar, total nitrogen, chlorine, protein, sugar/alkali, two sugar differences, and there is significant difference in potassium/chlorine, and other indicator differences are not remarkable.

Carry out further multiple ratio to existing between extremely significant each index:

From table, 36-38 can find out, there are utmost point significant difference in bottom oil content and reducing sugar between different aesthetic quality's score values, and there is significant difference in total reducing sugar; There are utmost point significant difference in middle part oil content, total reducing sugar, reducing sugar and sugar/alkali between different aesthetic quality's score values, and there are utmost point significant difference in top colourity, nicotine, total reducing sugar, reducing sugar, total nitrogen, protein and two sugar differences between different aesthetic quality's score values.

Bottom and top color are along with the reduction of the total points of smokeing panel test, thin out gradually, the middle part color deepens gradually, colourity etc. do not have obvious Changing Pattern, reduction along with the total points of smokeing panel test, under, in and top nicotine the trend of increase is arranged, each position total reducing sugar, reducing sugar and sugar/alkali content reduce thereupon, total nitrogen and protein increase.

2.6 the index screening of leading typical sample mass discrepancy and interval are just sentenced

2.6.1 index screening

Comprehensive above-mentioned many analyses, consider the inner link between quality factor, in conjunction with experience, we can filter out each physical and chemical index of Tobacco Leaves in Guizhou: surface density, nicotine, total reducing sugar, total nitrogen, potassium, sugar/alkali, two sugar ratio and potassium/chlorine can be used as the main physical and chemical index of its characteristic sound tobacco.

2.6.2 preferably sample is determined

Division according to ordered sample Fisher sorting technique to the quality of smokeing panel test, its division result is carried out to difference and smoke panel test total points scope sample physical and chemical index multivariate analysis of variance and single-factor test of hypothesis and further multiple ratio, according to central value ± SD(standard deviation) confidence level that drops in population sample of the numerical value interval of limiting is 68.2%, central value ± 1.96SD(standard deviation), be 95%, central value ± 2.57SD(standard deviation), be 99%.We are loosened to next segmentation by bottom and middle part through cutting apart the score value scope obtained, and the suitable quality of smokeing panel test in bottom is >=37.6 minutes, and the suitable quality of smokeing panel test in middle part is >=38.9 minutes.

2.6.3 sentence at the beginning of between main physical and chemical index Suitable Area

Comprehensive above Mathematical Statistics Analysis, we are judged at the beginning of between the main physical and chemical index the most suitable region of Tobacco Leaves in Guizhou:

Surface density: bottom 38.25-51.55g/m2, middle part 51.16-69.94g/m2, top 67.26-92.96g/m2; Nicotine: bottom 1.18-2.00%, middle part 1.46-2.58%, top 1.88-3.46%; Total reducing sugar: bottom 14.43-26.37%, middle part 24.26-31.08%, top 20.24-29.06%; Total nitrogen: bottom 1.70-2.30%, middle part 1.63-2.17%, top 1.62-2.50%; Potassium: bottom 1.65-2.55%, middle part 1.49-2.17%, top 1.07-1.73%; Sugar/alkali: bottom 6.49-18.65, middle part 8.60-19.26, top 4.53-15.63; Two sugar ratios: bottom 0.80-0.96, middle part 0.89-0.95, top 0.91-0.97; Potassium/chlorine: bottom 3.66-7.42, middle part 4.10-7.64, top 2.45-4.97.

The introducing of external sample and checking

We can be between the Tobacco Leaves in Guizhou the most suitable region by above-mentioned obtained section definition, now introduce external sample obtained interval is verified.Sampling before external sample is mainly derived from the bulk production sampling and beats leaf and rechecks, also comprise small part peasant household field sampling etc. simultaneously.The index screened is preferentially thought nicotine > sugar/alkali two sugar than potassium/chlorine verified.Be decided to be and take sample number 60% degree of conformity is restrictive condition.According to typical sample two sugar than and potassium chlorine than data characteristics, and two sugar are than with potassium chlorine, comparing the rule that affects of quality of tobacco, Guizhou, base, Shanghai characteristic high-quality tobacco two sugar than and potassium chlorine be judged to two sugar ratios>0.8, potassium chlorine ratio>2.5 at the beginning of between ratio Suitable Area.Now introducing external sample just sentences interval to obtained optimum and carries out the degree of conformity checking.

From degree of conformity, on the whole, the nicotine adopted, sugar/alkali, two sugar ratios and 4 indexs of potassium/chlorine (wherein two sugar ratios and potassium chlorine are than only limiting lower limit) are mainly manifested in the nicotine degree of conformity lower than 60%.Between each year, the degree of conformity of 2008 annual tobacco samples is the poorest, and bottom only has 27.50%.Obtain the suitable interval value of the main physical and chemical index of tobacco leaf, just sentencing on interval basis, must the synthesise various situation carry out concrete correction.

2.8 determining between the main physical and chemical index Suitable Area of tobacco leaf

The above analysis, according to the degree of conformity requirement, actual and Shanghai company raw tobacco material quality system routine chemical components harmony evaluation act.std (in Table 45) and other results of study in conjunction with Tobacco Leaves in Guizhou, and according to preferred sample frequency distribution situation, under the scope allowed in the suitable quality of smokeing panel test, we will affect each physical and chemical index optimum interval value of Tobacco Leaves in Guizhou and suitably revise, and with this, draw the suitable interval value of the main physical and chemical index of Tobacco Leaves in Guizhou (in Table 46).

According to resulting suitable interval value, again introduce external sample it is carried out to the degree of conformity checking.Because the 2008 annual sample emphasis that obtain are representative samples, outside sample limited amount, will not carry out the degree of conformity checking in introducing.

By 2007 years and the 2009 annual and sample detection results that gather in 2 years, according to the divided quality score of smokeing panel test, revised suitable interval value being quoted to 8 main physical and chemical indexs of tobacco leaf such as surface density, nicotine, total reducing sugar, total nitrogen, potassium, sugar/alkali, two sugar ratios and potassium/chlorine is verified, from the checking situation, can find out, except 2009 year middle part nicotine degrees of conformity, lower than 60%, all the other index degrees of conformity are all higher than 60%.The individual cases that occur may with sample to obtain quality relevant, degree of conformity is preferably the sample in 2007 years, this part sample cluster is for selecting sample, and the sample cluster in 2009 years is fully from sampling on the face in the tobacco purchasing process, corresponding hierarchical organization and purity are slightly weaker, thereby further illustrate suitability and the rationality of determined interval value.

2.9 the deduction of sound tobacco cultivation step

In 2007 years, the sample obtained has carried out determining plot, simultaneously for corresponding peasant household and plot, has carried out the investigation statistics of basic document, in order to infer the cultivation step of sound tobacco and ecological etc.According to obtained suitable interval value, corresponding corresponding sample cluster, reject the outer sample of interval value.We find, the quality of smokeing panel test of the tobacco leaf sample after rejecting is all better, and corresponding peasant household and plot, than system with to account for a certain proportion of be to wash Ma Xiang-Sun Wuyuan-tea La Shu He Gaotai town-Li Kuiwen-two of well banks plot.Two plot basic documents are as follows:

Table 48 and table 49, something in common is: all plantation is native cigarette, and soil fertility is medium, and the fertilizer of using is all special fertilizer for tobacco and topdresses, two places have all used fertilizer (oily gruel, cow dung and compound organic and inorganic fertilizer etc.) simultaneously, rational close planting and reasonably pinch and stay the number of sheets.Can draw, produce the raw tobacco material that meets our cigarette industry of high-quality, we can use for reference those measures, i.e. standard leaf tobacco production management, the standardized production level of raising flue-cured tobacco.According to above-mentioned measure, we can infer simultaneously, the production of sound tobacco, and it is crucial carrying out " in a cigarette " this piece of article.

Claims (4)

1. one kind quality of tobacco evaluation method based on the sensory evaluating smoking, it is characterized in that: the method with all improved final purpose of tobacco be all using the tobacco leaf sensory evaluating smoking well as target as starting point, take the tobacco leaf sensory evaluating smoking as basis, utilize simple correlation, partial correlation and successive Regression, the Double Selection successive Regression, path analysis, and the method such as artificial neural network is carried out preliminary assessment and screening to the indices that obtains sample, the main physical and chemical index that affects the tobacco leaf characteristic is tentatively determined, utilize ordered sample Fisher classification or Hierarchical Clustering method to be divided tobacco leaf sensory evaluating smoking quality, sample under the segmentation of cutting apart is carried out to assay and then determine the main physical and chemical index that affects the tobacco leaf characteristic, carry out determining of optimum sample simultaneously, and according to sample number>60% degree of conformity is restrictive condition, obtain and affect the main physical and chemical index optimum range of tobacco leaf sensory evaluating smoking, and instruct the buying of tobacco industrial enterprise raw tobacco material with utilization and according to main physical and chemical index optimum range according to main physical and chemical index and optimum range thereof, the tobacco leaf sample that reverse find meets the demands, continuous and draw the ecologic environment of this sample growth, kind and cultivation technique and the measure that forms a complete production network, instruct tobacco commercial enterprise leaf tobacco production.

2. according to claim 1 Quality of tobacco evaluation method based on the sensory evaluating smoking, it is characterized in that: the method comprises the following steps:

(1) sample acquisition, require institute's collecting sample and grade contain wide, purity is high, representativeness is strong;

(2) pattern detection, obtain sample data by sensory evaluating smoking, industrial applicability analysis, mensuration presentation quality, physical characteristics and all kinds of chemical compositions of profile sample and the relevant index of quality;

(3) former data are processed, and step (2) the data obtained resource is arranged, and make it to be beneficial to statistical study, extrapolate sugared alkali ratio, two sugar ratios, potassium chlorine ratio, two sugar difference and all kinds of indexs that affect cigarette quality, the rejecting abnormalities value;

(4) analyzed according to the actual conditions of step (3) gained sample data and sample sampling spot, guaranteed that sample data meets representativeness, wide in range property and discreteness;

(5) evaluation index primary election, because all improved final purpose of tobacco is all that to take the tobacco leaf sensory evaluating smoking be well target, so take the tobacco leaf sensory evaluating smoking as basis, utilize simple correlation, partial correlation and successive Regression, Double Selection successive Regression, path analysis, and the method such as artificial neural network carries out preliminary assessment and screening to the quality evaluation index in step (3), preliminary screening goes out to affect the main physical and chemical index of tobacco leaf characteristic;

(6), on the basis of step (5), utilize ordered sample Fisher classification or Hierarchical Clustering method to the smoked panel test division of quality of sample; A utilizes the test of hypothesis of multivariate analysis of variance single-factor and multiple ratio by the affiliated sample of the segmentation of cutting apart, and further the screening and assessment index, determine relatively convenient, accurate and can reflect tobacco style and quality stability control index; The sample decomposition that b carries out tobacco leaf sensory evaluating smoking quality according to methods such as ordered sample Fisher classification or Hierarchical Clusterings, in conjunction with central value ± SD(standard deviation), the confidence level that the numerical value interval of limiting is dropped in population sample is 68.2%, central value ± 1.96SD(standard deviation) be 95%, central value ± 2.57SD(standard deviation) be 99% requirement, determine optimum sample sensory evaluating smoking scope, then determine optimum sample cluster;

(7) verification and debugging optimum range

According to sample number > 60% degree of conformity is restrictive condition, in conjunction with cigarette enterprise to the raw tobacco material quality requirements, introduce the external certificate sample, carry out the degree of conformity checking according to determined main physical and chemical index in step (6), and then debugging and revision index of correlation parameter, obtain the main physical and chemical index optimum range of tobacco leaf;

(8) instruct the buying of tobacco industrial enterprise raw tobacco material and utilize

The main physical and chemical index of tobacco leaf and optimum range thereof according to obtaining in step (7), carry out tobacco leaf classification and purchase, instructs the buying of tobacco industrial enterprise raw tobacco material and utilize;

(9) reckoning of ideotype growing way appearance and the measure that forms a complete production network

According to the main physical and chemical index of tobacco leaf and the optimum range thereof that obtain in step (7), the tobacco leaf sample that reverse find meets the demands, continue and draw ecologic environment, kind and cultivation technique and the measure that forms a complete production network that this sample is grown, instructing tobacco commercial enterprise leaf tobacco production.

3. according to claim 2 Quality of tobacco evaluation method based on the sensory evaluating smoking, it is characterized in that: in step (1), sample divides typical sample and outside two parts of checking sample of introducing, and when obtaining sample, correspondence is carried out investigation the record of sampling spot ecology, kind and cultivation technique;

The characteristics such as that the selected grade that requires a, typical sample contains is wide, purity is high, representative strong, this part sample detects the parameter of obtaining will be as basic assay data;

B, the outside checking sample of introducing, require originate wide, wide ranges, can represent that Cigarette Industrial Enterprise raw tobacco material buying level gets final product, and this part sample detects the parameter of obtaining will or revise data as checking.

4. the quality of tobacco evaluation method based on the sensory evaluating smoking according to claim 2 is characterized in that: in step (5), preliminary screening goes out to affect the main physical and chemical index of tobacco leaf characteristic and is: surface density, nicotine, total reducing sugar, total nitrogen, potassium, sugar/alkali, two sugar than and potassium/chlorine.

Images