CN103674884A - Random forest classification method for tobacco leaf style characteristics based on near infrared spectral information - Google Patents
Random forest classification method for tobacco leaf style characteristics based on near infrared spectral information Download PDFInfo
- Publication number
- CN103674884A CN103674884A CN201210344642.4A CN201210344642A CN103674884A CN 103674884 A CN103674884 A CN 103674884A CN 201210344642 A CN201210344642 A CN 201210344642A CN 103674884 A CN103674884 A CN 103674884A
- Authority
- CN
- China
- Prior art keywords
- sample
- near infrared
- spectrum
- tobacco
- random forest
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Abstract
The invention discloses a random forest classification method for tobacco leaf style characteristics based on near infrared spectral information. The random forest classification method is characterized by being implemented by the following steps of (1) performing modeling sample preparation; (2) performing spectrum scanning; (3) performing spectrum preprocessing; (4) exporting spectrum data; (5) performing modeling; (6) scanning a sample to be detected according to the steps (2)-(4), obtaining near infrared spectrum data of the sample to be detected, and respectively comparing the near infrared spectrum data with a math model in a memory to obtain factors, such as a planting region, the species and the grade part, of tobacco leaves, which are used for determining the style characteristics. According to the random forest classification method, the near infrared spectrum information of the tobacco leaves is used as an object, and the planting region, the species and the grade part which are used for determining the tobacco leaf style characteristics are subjected to mode identification by the random forest classification method. The random forest classification method disclosed by the invention is easy to operate, time-saving and labor-saving; the sample does not need to be subjected to pretreatment; the information such as the planting region, the species and the grade part of the sample can be obtained within 2 minutes by directly scanning a near infrared spectrum.
Description
The invention belongs to a kind of sorting technique of tobacco style feature, is specifically that to take the near infrared light spectrum information of tobacco leaf be object, adopts random forest classification method to carry out pattern-recognition to determining planting area, kind, the grade position of tobacco style feature.
Technical background
Raw tobacco material is core and the basis of cigarette composition, and for guaranteeing the stable of cigarette composition, the formula personnel of cigarette enterprise adopt many grades, multizone, eurypalynous tobacco compatibility to form the tobacco leaf formulation of cigarette.According to the feature of Chinese-style cigarette, the style and features of the formation of cigarette product style characteristic based on tobacco leaf.
Style and features is the soul of cigarette product, and tobacco style feature is again that cigarette product forms the basis of style and features and ensures.Statement as tobacco style feature does not also have generally accepted standard, with the quantitative measurement result of physics and chemistry, comes the style and features of definite statement tobacco leaf also very difficult.The factor that affects tobacco style feature mainly contains: ecologic environment, kind and cultivation modulation technique, the factor of characteristic sound tobacco development project based on affecting tobacco style feature set out, and plants the tobacco leaf of selected kind in certain region by specific cultivation modulation technique.Cigarette enterprise, in reality purchase tobacco leaf and tobacco leaf formulation formulation process, is mainly considered three factors such as planting area, kind, grade position.Based on this, we think that planting area, kind, grade position have determined the style and features of tobacco leaf.
Its chemical composition of tobacco leaf at different planting areas, kind, grade position there are differences, chromatographic fingerprinting has the advantages that fingerprint characteristic is analyzed, can carry out qualitative and quantitative analysis to micro-trace constituent, be applicable to Analysis of Complex composing system, being the method that current tobacco leaf trace trace chemistry constituent analysis evaluation generally adopts, is the focus of research at present by the different of its chemical composition relatively and then the tobacco leaf of identifying different planting areas, kind, grade position.But chromatogram chromatographic technique focuses on the compartment analysis for aroma component, isolated globality, harmony and the synergy of essence and flavoring agent, and chromatographic technique pretreatment process is numerous and diverse, human and material resources, time cost are higher.Near infrared spectrum is mainly that the anharmonicity due to molecular vibration produces while making molecular vibration from ground state to high level transition, what record is frequency multiplication and the sum of fundamental frequencies absorption of hydric group vibration, utilize the group features stronger to near-infrared absorption such as C-H in material, N-H, O-H and C=O, according to the near infrared light spectrum information of organic substance in conjunction with Chemical Measurement to corresponding composition or index carry out quantitatively, observational measurement.The related information of the tobacco components that near infrared spectrum comprises is very abundant, based on Near Infrared Information, carry out tobacco leaf cluster analysis and pattern-recognition has reliable material base, the qualitative, quantitative research that application Near Infrared Information carries out quality of tobacco has broad application prospects.Acquisition with respect to micro-trace chemistry composition, near-infrared spectrum technique has huge advantage, being obtained by near infrared spectrometer scanning of near infrared spectrogram, procurement process speed is fast, sample does not need pre-service, simple to operate, personnel require low, without waste, pollution-free.
Random forest is a kind of assembled classifier method being proposed in calendar year 2001 by Leo Breiman and Adele Cutler.The fundamental classifier that forms random forest is called decision tree, and the classification of its output is the mode of the classification exported by indivedual trees.
Current near infrared light spectrum information of take tobacco leaf is object, adopts random forest classification method to carry out the method for pattern-recognition and have not been reported determining planting area, kind, the grade position of tobacco style feature.
Summary of the invention
For the weak point of present technology existence, the invention provides a kind of random forest classification method of the tobacco style feature based near infrared light spectrum information.The method does not need sample to carry out pre-treatment, easy and simple to handle, and analytical test speed is fast, and measurement result is accurate, reproducible.
In order to solve the problems of the technologies described above, the present invention adopts following technical scheme:
A random forest classification method for tobacco style feature based near infrared light spectrum information, the method realizes by following steps:
(1) modeling sample is prepared: collect and obtain tobacco sample, tobacco sample state be pipe tobacco or powder all can, sample need comprise the information such as planting area, grade position, kind;
(2) spectral scan: obtain its near infrared spectrogram by near infrared spectrometer scanning modeling sample, the running parameter of instrument is: spectral range 10000~3800cm
-1, resolution 4~32cm
-1, scanning and be averaged spectrum 1~100 time, each Sample Scan obtains 2 above averaged spectrum;
(3) spectrum pre-service: adopt standard canonical transformation to eliminate the inhomogeneous difference of bringing of sample, adopt the level and smooth spectrum of Norris's smothing filtering, eliminate the low-frequency information that high frequency noise remains with use, employing second-order differential is processed, eliminate the impact of baseline wander, obtain than more high resolving power and the spectral profile variation more clearly of former spectrum;
(4) derive spectroscopic data: by 10000~3800cm
-1the data point of spectral range derives; Also spectrum can be carried out after major component is processed deriving its principal component scores data;
(5) set up model: the near-infrared spectra diagram data of modeling sample is combined with sample style and features information, and modeling sample is divided into 10~90% at random, part sample is for random forest modeling, other sample is for modelling verification, decision tree is more than 50, Nodes variable number is more than 2, repeatedly, Optimized model progressively, make it reach optimum condition, the final classification of modeling sample is determined by other mode of output class, provides the accuracy that in modeling sample training process, classification is differentiated simultaneously.The disaggregated model that foundation is obtained deposits in storer;
(6) sample detection: by step 2~4 scanning testing sample, after processing, obtain its near infrared spectrum data, respectively with storer in compare planting area, kind, grade position etc. that can obtain tobacco leaf of mathematical model determine the factor of style and features.
The usefulness of technique scheme is:
It is object that the near infrared light spectrum information of tobacco leaf is take in the present invention, adopts random forest classification method to carry out pattern-recognition to determining planting area, kind, the grade position of tobacco style feature.The present invention operation is easier, time saving and energy saving, does not need sample to carry out pre-treatment, directly scans the information such as planting area that its near infrared spectrogram just can obtain sample in 2 minutes, kind, grade position.
Accompanying drawing explanation
Nothing
Embodiment
Below in conjunction with specific embodiment, further set forth the present invention.These embodiment are only not used in and limit the scope of the invention for the present invention is described.
Embodiment 1:
A random forest classification method for tobacco style feature based near infrared light spectrum information, the method realizes by following steps:
(1) modeling sample is prepared: collect and obtain 838 powder tobacco samples, the planting area of sample is respectively tobacco planting district, the southeast, southwestern tobacco planting district, the Yellow River and Huai He River tobacco planting district, Middle And Upper Reaches of The Yangtze River tobacco planting district;
(2) spectral scan: obtain its near infrared spectrogram by near infrared spectrometer scanning modeling sample, the running parameter of instrument is: spectral range 8500~4000cm
-1, resolution 8cm
-1, scanning and be averaged spectrum 20 times, each Sample Scan obtains 5 averaged spectrum;
(3) spectrum pre-service: adopt standard canonical transformation to eliminate the inhomogeneous difference of bringing of sample, adopt the level and smooth spectrum of Norris's smothing filtering, eliminate the low-frequency information that high frequency noise remains with use, employing second-order differential is processed, eliminate the impact of baseline wander, obtain than more high resolving power and the spectral profile variation more clearly of former spectrum.
(4) derive spectroscopic data: will
8500~4000cm -1 the data point of spectral range derives, and every spectrum obtains 1038 data points;
(5) set up model: the near infrared spectrum data of modeling sample is combined with sample planting area information, and modeling sample is divided into 30% at random, part sample is for random forest modeling, and other sample is for modelling verification, and decision tree is 100, Nodes variable number is 32, repeatedly, progressively Optimized model, makes it reach optimum condition, the final classification of modeling sample is determined by other mode of output class, provides the accuracy that in modeling sample training process, classification is differentiated simultaneously.The tobacco leaf planting territorial classification model that foundation is obtained deposits in storer;
(6) sample detection: the testing sample of 100 different planting areas obtains its near infrared spectrum data after processing by step 2~4, with the compare information of the planting area that can obtain tobacco leaf of the tobacco leaf planting region mathematical model in storer.
The accuracy of modeling sample is 93.32%, and the accuracy of sample detection is at 84%(table 1).
Table 1 tobacco planting district random forest classification mode recognition result
Embodiment 2:
A random forest classification method for tobacco style feature based near infrared light spectrum information, the method realizes by following steps:
(1) modeling sample is prepared: collect and obtain 647 powder tobacco samples, the kind of sample is respectively cloud and mist 87, cloud and mist 85, dark green-1, K326, the large gold dollar of safflower, F1-35;
(2) spectral scan: obtain its near infrared spectrogram by near infrared spectrometer scanning modeling sample, the running parameter of instrument is: spectral range 9000~3800cm
-1, resolution 16cm
-1, scanning and be averaged spectrum 50 times, each Sample Scan obtains 10 averaged spectrum;
(3) spectrum pre-service: adopt standard canonical transformation to eliminate the inhomogeneous difference of bringing of sample, adopt the level and smooth spectrum of Norris's smothing filtering, eliminate the low-frequency information that high frequency noise remains with use, employing second-order differential is processed, eliminate the impact of baseline wander, obtain than more high resolving power and the spectral profile variation more clearly of former spectrum.
(4) derive spectroscopic data: spectrum is carried out after major component is processed deriving its front 10 principal component scores data.
(5) set up model: the near infrared spectrum principal component scores data of modeling sample are combined with the kind information of sample, and modeling sample is divided into 50% at random, part sample is for random forest modeling, other sample is for modelling verification, decision tree is 200, Nodes variable number is 4, repeatedly, Optimized model progressively, make it reach optimum condition, the final classification of modeling sample is determined by other mode of output class, provides the accuracy that in modeling sample training process, classification is differentiated simultaneously.The tobacco leaf assortment model that foundation is obtained deposits in storer;
(6) sample detection: the testing sample of 90 different cultivars obtains its near infrared spectrum principal component scores data after processing by step 2~4, compares and can obtain the kind information of tobacco leaf with the tobacco leaf kind mathematical model in storer.
Because the main cultivation flue-cured tobacco cultivars of current China all directly or indirectly comes from identical parent, even between some kind, sibship is extremely near, while causing kind to be differentiated, easily misjudges.Adopt the accuracy of Method Modeling sample of the present invention still to reach 81.41%, the accuracy of sample detection has reached 64.44%(table 2).
Table 2 tobacco bred machine forest classified pattern-recognition result
Embodiment 3:
A random forest classification method for tobacco style feature based near infrared light spectrum information, the method realizes by following steps:
(1) modeling sample is prepared: collect and obtain 832 tobacco sample, the grade position of sample is respectively B2F C3F X2F;
(2) spectral scan: obtain its near infrared spectrogram by near infrared spectrometer scanning modeling sample, the running parameter of instrument is: spectral range 10000~4500cm
-1, resolution 32cm
-1, scanning and be averaged spectrum 80 times, each Sample Scan obtains 20 averaged spectrum;
(3) spectrum pre-service: adopt standard canonical transformation to eliminate the inhomogeneous difference of bringing of sample, adopt the level and smooth spectrum of Norris's smothing filtering, eliminate the low-frequency information that high frequency noise remains with use, employing second-order differential is processed, eliminate the impact of baseline wander, obtain than more high resolving power and the spectral profile variation more clearly of former spectrum.
(4) derive spectroscopic data: spectrum is carried out after major component is processed deriving its front 20 principal component scores data.
(5) set up model: the near infrared spectrum principal component scores data of modeling sample are combined with the grade location information of sample, and modeling sample is divided into 80% at random, part sample is for random forest modeling, other sample is for modelling verification, decision tree is 300, Nodes variable number is 6, repeatedly, Optimized model progressively, make it reach optimum condition, the final classification of modeling sample is determined by other mode of output class, provides the accuracy that in modeling sample training process, classification is differentiated simultaneously.The tobacco leaf grade position disaggregated model that foundation is obtained deposits in storer;
(6) sample detection: the testing sample of 100 different cultivars obtains its near infrared spectrum principal component scores data after processing by step 2~4, compares and can obtain the grade location information of tobacco leaf with the tobacco leaf grade position mathematical model in storer.
Adopt the accuracy of Method Modeling sample of the present invention to reach 90.24%, the accuracy of sample detection has reached 85.00%(table 3).
Table 3 tobacco grade position machine forest classified pattern-recognition result
Above are only three specific embodiments of the present invention, but design concept of the present invention is not limited to this, allly utilizes this design to carry out the change of unsubstantiality to the present invention, all should belong to the behavior of invading protection domain of the present invention.
Claims (2)
1. the random forest classification method of the tobacco style feature based near infrared light spectrum information, the method realizes by following steps:
(1) modeling sample is prepared: collect and obtain tobacco sample, tobacco sample state be pipe tobacco or powder all can, sample need comprise the information such as planting area, grade position, kind;
(2) spectral scan: obtain its near infrared spectrogram by near infrared spectrometer scanning modeling sample, the running parameter of instrument is: spectral range 10000~3800cm
-1, resolution 4~32cm
-1, scanning and be averaged spectrum 1~100 time, each Sample Scan obtains 2 above averaged spectrum;
(3) spectrum pre-service: adopt standard canonical transformation to eliminate the inhomogeneous difference of bringing of sample, adopt the level and smooth spectrum of Norris's smothing filtering, eliminate the low-frequency information that high frequency noise remains with use, employing second-order differential is processed, eliminate the impact of baseline wander, obtain than more high resolving power and the spectral profile variation more clearly of former spectrum;
(4) derive spectroscopic data: by 10000~3800cm
-1the data point of spectral range derives;
(5) set up model: the near-infrared spectra diagram data of modeling sample is combined with sample style and features information, and modeling sample is divided into 10~90% at random, part sample is for random forest modeling, other sample is for modelling verification, decision tree is more than 50, Nodes variable number is more than 2, repeatedly, Optimized model progressively, make it reach optimum condition, the final classification of modeling sample is determined by other mode of output class, provides the accuracy that in modeling sample training process, classification is differentiated simultaneously.The disaggregated model that foundation is obtained deposits in storer;
(6) sample detection: by step 2~4 scanning testing sample, after processing, obtain its near infrared spectrum data, respectively with storer in compare planting area, kind, grade position etc. that can obtain tobacco leaf of mathematical model determine the factor of style and features.
2. the method for the auxiliary cigarette composition of a kind of SIMCA based near infrared light spectrum information as claimed in claim 1, is characterized in that: step 4 is replaceable one-tenth also: spectrum is carried out after major component is processed deriving its principal component scores data.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201210344642.4A CN103674884A (en) | 2012-09-17 | 2012-09-17 | Random forest classification method for tobacco leaf style characteristics based on near infrared spectral information |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201210344642.4A CN103674884A (en) | 2012-09-17 | 2012-09-17 | Random forest classification method for tobacco leaf style characteristics based on near infrared spectral information |
Publications (1)
Publication Number | Publication Date |
---|---|
CN103674884A true CN103674884A (en) | 2014-03-26 |
Family
ID=50313094
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201210344642.4A Pending CN103674884A (en) | 2012-09-17 | 2012-09-17 | Random forest classification method for tobacco leaf style characteristics based on near infrared spectral information |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103674884A (en) |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104568823A (en) * | 2015-01-07 | 2015-04-29 | 中国农业大学 | Tobacco leaf raw material proportioning ratio calculation method and tobacco leaf raw material proportioning ratio calculation device based on near infrared spectrum |
CN104899551A (en) * | 2015-04-30 | 2015-09-09 | 北京大学 | Form image classification method |
CN105891147A (en) * | 2016-03-30 | 2016-08-24 | 浙江中烟工业有限责任公司 | Near infrared spectrum information extraction method based on canonical correlation coefficients |
CN106501208A (en) * | 2016-09-20 | 2017-03-15 | 广西中烟工业有限责任公司 | A kind of tobacco style similitude sorting technique based near infrared light spectrum signature |
CN109374575A (en) * | 2018-12-20 | 2019-02-22 | 中国烟草总公司郑州烟草研究院 | A kind of discrimination method of the cured tobacco leaf background color based on near-infrared spectral analysis technology |
CN110363119A (en) * | 2019-06-28 | 2019-10-22 | 广东中烟工业有限责任公司 | Based on wavelet transformation-random forests algorithm tobacco mildew method for quickly identifying |
CN110907379A (en) * | 2019-12-23 | 2020-03-24 | 鞍钢集团矿业有限公司 | Method for measuring components and content of Anshan type iron ore based on random forest algorithm |
CN110987865A (en) * | 2019-12-13 | 2020-04-10 | 齐鲁工业大学 | Method for detecting fig quality based on near infrared spectrum |
CN111624265A (en) * | 2020-04-22 | 2020-09-04 | 南京农业大学 | Method for identifying egg species |
CN111735807A (en) * | 2020-07-14 | 2020-10-02 | 征图新视(江苏)科技股份有限公司 | LIBS-based online detection method for large-leaf crop parts and oil content |
CN112418659A (en) * | 2020-11-20 | 2021-02-26 | 上海魔橙网络科技有限公司 | Cross-border commodity tracing method, system and device based on block chain system |
CN113124784A (en) * | 2021-03-17 | 2021-07-16 | 红塔烟草(集团)有限责任公司 | Primary flue-cured tobacco part judgment method based on laser scanning confocal technology |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1831516A (en) * | 2006-04-03 | 2006-09-13 | 浙江大学 | Method for nondistructive discriminating variety and true and false of cigarette using visible light and near-infrared spectrum technology |
CN1982874A (en) * | 2005-12-16 | 2007-06-20 | 天津天士力制药股份有限公司 | Near-infrared diffuse reflection spectral method for fastly inspecting drop effective ingredient content |
CN101101260A (en) * | 2007-05-25 | 2008-01-09 | 吉林大学 | Near infrared spectrum damage-free analysis method for anti-tuberculosis drugs |
CN101251471A (en) * | 2008-03-12 | 2008-08-27 | 湖南中烟工业有限责任公司 | Method for searching analog tobacco leaf based on tobacco leaf near infrared spectra |
CN101303294A (en) * | 2008-06-20 | 2008-11-12 | 河南中医学院 | Application method of near-infrared on-line test technology in Chinese medicine Yiqing granule production |
CN101995388A (en) * | 2009-08-26 | 2011-03-30 | 北京凯元盛世科技发展有限责任公司 | Near infrared quality control analysis method and system of tobacco |
CN102374976A (en) * | 2011-10-21 | 2012-03-14 | 中国兵器工业集团第五三研究所 | Correlation model for detecting SBR (styrene butadiene rubber) content of compounded rubber |
-
2012
- 2012-09-17 CN CN201210344642.4A patent/CN103674884A/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1982874A (en) * | 2005-12-16 | 2007-06-20 | 天津天士力制药股份有限公司 | Near-infrared diffuse reflection spectral method for fastly inspecting drop effective ingredient content |
CN1831516A (en) * | 2006-04-03 | 2006-09-13 | 浙江大学 | Method for nondistructive discriminating variety and true and false of cigarette using visible light and near-infrared spectrum technology |
CN101101260A (en) * | 2007-05-25 | 2008-01-09 | 吉林大学 | Near infrared spectrum damage-free analysis method for anti-tuberculosis drugs |
CN101251471A (en) * | 2008-03-12 | 2008-08-27 | 湖南中烟工业有限责任公司 | Method for searching analog tobacco leaf based on tobacco leaf near infrared spectra |
CN101303294A (en) * | 2008-06-20 | 2008-11-12 | 河南中医学院 | Application method of near-infrared on-line test technology in Chinese medicine Yiqing granule production |
CN101995388A (en) * | 2009-08-26 | 2011-03-30 | 北京凯元盛世科技发展有限责任公司 | Near infrared quality control analysis method and system of tobacco |
CN102374976A (en) * | 2011-10-21 | 2012-03-14 | 中国兵器工业集团第五三研究所 | Correlation model for detecting SBR (styrene butadiene rubber) content of compounded rubber |
Non-Patent Citations (5)
Title |
---|
XIAOHUI LIN ET AL: "A random forest of combined features in the classification of cut tobacco based on gas chromatography fingerprinting", 《TALANTA》, vol. 82, 30 July 2010 (2010-07-30), XP027241627 * |
杜文等: "基于近红外光谱的烟叶SIMCA模式识别", 《中国烟草学报》, vol. 15, no. 5, 31 October 2009 (2009-10-31) * |
王遵义等: "基于小波变换和神经网络的可见-近红外光谱对烟草品种的鉴别", 《浙江大学学报(农业与生命科学版)》, vol. 35, no. 6, 31 December 2009 (2009-12-31) * |
章英等: "基于LS-SVM 的烤烟烟叶产地判别", 《湖北农业科学》, vol. 51, no. 3, 29 February 2012 (2012-02-29) * |
鲍峰伟: "近红外光谱分析技术在烟草化学指标定量分析中的应用", 《中国优秀硕士学位论文全文数据库工程科技Ⅰ辑》, no. 4, 15 October 2007 (2007-10-15) * |
Cited By (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104568823A (en) * | 2015-01-07 | 2015-04-29 | 中国农业大学 | Tobacco leaf raw material proportioning ratio calculation method and tobacco leaf raw material proportioning ratio calculation device based on near infrared spectrum |
CN104568823B (en) * | 2015-01-07 | 2017-08-04 | 中国农业大学 | Raw tobacco material ratio computational methods and device based near infrared spectrum |
CN104899551A (en) * | 2015-04-30 | 2015-09-09 | 北京大学 | Form image classification method |
CN104899551B (en) * | 2015-04-30 | 2018-08-14 | 北京大学 | A kind of form image sorting technique |
CN105891147A (en) * | 2016-03-30 | 2016-08-24 | 浙江中烟工业有限责任公司 | Near infrared spectrum information extraction method based on canonical correlation coefficients |
CN106501208A (en) * | 2016-09-20 | 2017-03-15 | 广西中烟工业有限责任公司 | A kind of tobacco style similitude sorting technique based near infrared light spectrum signature |
CN109374575A (en) * | 2018-12-20 | 2019-02-22 | 中国烟草总公司郑州烟草研究院 | A kind of discrimination method of the cured tobacco leaf background color based on near-infrared spectral analysis technology |
CN110363119A (en) * | 2019-06-28 | 2019-10-22 | 广东中烟工业有限责任公司 | Based on wavelet transformation-random forests algorithm tobacco mildew method for quickly identifying |
CN110987865A (en) * | 2019-12-13 | 2020-04-10 | 齐鲁工业大学 | Method for detecting fig quality based on near infrared spectrum |
CN110907379A (en) * | 2019-12-23 | 2020-03-24 | 鞍钢集团矿业有限公司 | Method for measuring components and content of Anshan type iron ore based on random forest algorithm |
CN111624265A (en) * | 2020-04-22 | 2020-09-04 | 南京农业大学 | Method for identifying egg species |
CN111735807A (en) * | 2020-07-14 | 2020-10-02 | 征图新视(江苏)科技股份有限公司 | LIBS-based online detection method for large-leaf crop parts and oil content |
CN112418659A (en) * | 2020-11-20 | 2021-02-26 | 上海魔橙网络科技有限公司 | Cross-border commodity tracing method, system and device based on block chain system |
CN113124784A (en) * | 2021-03-17 | 2021-07-16 | 红塔烟草(集团)有限责任公司 | Primary flue-cured tobacco part judgment method based on laser scanning confocal technology |
CN113124784B (en) * | 2021-03-17 | 2022-08-05 | 红塔烟草(集团)有限责任公司 | Primary flue-cured tobacco part judgment method based on laser scanning confocal technology |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103674884A (en) | Random forest classification method for tobacco leaf style characteristics based on near infrared spectral information | |
CN102866127B (en) | A kind of SIMCA based near infrared light spectrum information assists the method for cigarette composition | |
CN102023137B (en) | Method for identifying white spirits | |
Ren et al. | Multi-variable selection strategy based on near-infrared spectra for the rapid description of dianhong black tea quality | |
CN101413883B (en) | Method for identifying tea-leaf origin by infrared spectrum | |
CN102778442A (en) | Method for rapidly identifying type of balsam material liquid for cigarette | |
CN105806824A (en) | Quick edible oil variety identification method based on multisource spectral feature fusion | |
CN107796782A (en) | Redrying quality stability evaluation method based on tobacco leaf characteristic spectrum consistency metric | |
CN103411906B (en) | The near infrared spectrum qualitative identification method of pearl powder and oyster shell whiting | |
CN102879353A (en) | Near infrared detection method for contents of protein components in peanut | |
CN103487411A (en) | Method for recognizing steel grade by combining random forest algorithm with laser-induced breakdown spectroscopy | |
CN105136738A (en) | Near-infrared-based method for identifying tree varieties ranging from eucalyptus-category tree varieties to acacia-mangium-category tree varieties | |
CN106501208A (en) | A kind of tobacco style similitude sorting technique based near infrared light spectrum signature | |
CN105717066A (en) | Near-infrared spectrum recognition model based on weighting association coefficients | |
CN102937575B (en) | Watermelon sugar degree rapid modeling method based on secondary spectrum recombination | |
CN110749565A (en) | Method for rapidly identifying storage years of Pu' er tea | |
CN104020128A (en) | Method for rapidly identifying propolis source | |
CN105138834A (en) | Tobacco chemical value quantifying method based on near-infrared spectrum wave number K-means clustering | |
CN105181761A (en) | Method for rapidly identifying irradiation absorbed dose of tea by using electronic nose | |
CN105758819A (en) | Method for detecting organic components of soil by utilizing near infrared spectrum | |
CN102507829B (en) | Method for establishing tobacco HPLC (High Performance Liquid Chromatography) fingerprint database and application thereof | |
CN111523587A (en) | Woody plant species spectrum identification method based on machine learning | |
CN107632010A (en) | A kind of quantitative approach of combination LIBS to steel samples | |
CN104596976A (en) | Method for determining protein of paper-making reconstituted tobacco through ear infrared reflectance spectroscopy technique | |
CN102680427A (en) | Method for identifying surface fragrance quality of cigarettes by near infrared spectral analysis technology |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20140326 |
|
RJ01 | Rejection of invention patent application after publication |