CN102283243A - Solid bactericide pesticide classification method - Google Patents
Solid bactericide pesticide classification method Download PDFInfo
- Publication number
- CN102283243A CN102283243A CN201110178670A CN201110178670A CN102283243A CN 102283243 A CN102283243 A CN 102283243A CN 201110178670 A CN201110178670 A CN 201110178670A CN 201110178670 A CN201110178670 A CN 201110178670A CN 102283243 A CN102283243 A CN 102283243A
- Authority
- CN
- China
- Prior art keywords
- bactericide
- sample
- absorption
- utilize
- principal component
- 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
Images
Abstract
The invention discloses a solid bactericide pesticide classification method. The method comprises the following steps of: 1) detecting raw solid bactericide agents with purity of more than 98 percent by using a terahertz time domain system, acquiring time domain signals of samples, and calculating absorption spectrums, wherein the raw bactericide agents comprise ferbam, thiophanatemethyl, probenazole and captan; 2) analyzing the absorption spectrums of the samples by using a main component analysis method, and extracting features for classifying the spectrums of the samples; and 3) identifying the terahertz spectrums of the bactericide samples by using a radial basis function neural network. The terahertz time domain spectrum technology-based bactericide classification method is simple and quick in the whole experiment process and high in identification accuracy, is suitable for classification of various solid bactericide pesticides, and has certain market application prospect.
Description
Technical field
The present invention relates to the agricultural chemicals sorting technique, relate in particular to a kind of bactericide solid pesticide sorting technique.
Background technology
Along with modern agricultural development, a large amount of agricultural chemicals is applied in the agricultural production, and the application of agricultural chemicals plays an important role for the raising of crop pest control and crop yield.Yet, along with development of science and technology, constantly the introducing to the market of novel agrochemical, this has caused difficulty with regard to the classification of giving agricultural chemicals, only depends on vision and sense of smell to be difficult to differentiate, so a kind of sorting technique fast and effectively has great significance.
The terahertz light spectral technology is widely used in research fields such as food security, dangerous material detections in development recent years fast.The terahertz light spectral technology has the not available advantage of other spectrum, at first, the rotational energy level and the intermolecular force of a lot of materials all drop in the frequency band of tera-hertz spectra, these substance characteristics are that other spectrum does not observe, the tera-hertz spectra of every kind of material all is the reaction of its molecular structure, so also be the dactylogram of this material.Secondly the signal to noise ratio height of tera-hertz spectra is subjected to noise jamming little, and last, the Terahertz system is simply quick to the testing process of sample.
Principal component method is to use more a kind of feature extraction and sorting technique in recent years, it utilizes the thought of dimensionality reduction, the more data set of dimension is converted into less " effectively " feature of dimension to be represented, and the internal information content that does not reduce initial data and comprised, when sample being discerned with cluster, amount of calculation just must reduce greatly like this.
Artificial neural network is made up of a large amount of simple neurons, is used to simulate the network system of human brain thinking, can be used for solving complicated and nonlinear problem, has stronger self-organizing, self adaptation and study, association, fault-tolerant and antijamming capability.The formation of basic radial base neural net comprises three layers, and wherein each layer all has diverse effect.Input layer is by some source points, and promptly the perception unit is formed, and they link up network and external environment.The second layer is only hidden layer in the network, and its effect is to carry out nonlinear transformation between from the input space to the concealed space; In most of the cases concealed space has higher dimension.Output layer is linear, and it provides response for the enable mode that acts on input layer.Radial base neural net is a kind of of neutral net, and it adopts radially basic transfer function, relative and BP neutral net, and training speed is faster, and the effect of identification is also better.
This invention combines terahertz time-domain spectroscopic technology, principal component analysis method and neutral net, has realized the identification to solid bactericide class agricultural chemicals, is a kind of discrimination method of agricultural chemicals simply fast.
Summary of the invention
The objective of the invention is to overcome the deficiencies in the prior art, a kind of bactericide solid pesticide sorting technique is provided.
The step of bactericide solid pesticide sorting technique is as follows:
1) utilize the terahertz time-domain system that purity is detected greater than the former medicine of 98% solid bactericide, obtain the time-domain signal of each sample, and calculate absorption spectra, bactericide original drug is ferbam, thiophanate-methyl, probenazole and captan;
2) utilize principal component analysis method that the absorption spectra of each sample is analyzed, extract feature, be used for the classification of each sample spectra;
3) utilize radial base neural net that the tera-hertz spectra of each bactericide sample is discerned.
Described step 1) is:
(1) dry in drying box respectively purity greater than 98% Coromate, methyl sulfur bacterium bactericide, probenazole bactericide and captan bactericide original drug and polyethylene powders, in mortar, grind, then four kinds of bactericide original drugs are mixed by 1:1 with polyethylene respectively, depress to the bactericide sample of thickness with tablet press machine at the pressure of 30MPa at 1.3mm, the quality of bactericide sample is 200mg, and diameter is 13mm;
(2) humidity less than 4% nitrogen environment in, utilize the terahertz time-domain system that empty specimen holder and the specimen holder that is placed with the bactericide sample are detected, each duplicate detection 10 times obtains empty specimen holder reference signal and the specimen holder sample signal that is placed with the bactericide sample respectively; Utilize Fourier transformation to convert reference signal and sample signal to frequency domain, calculate the absorption spectra of each bactericide sample with fresnel formula, formula is as follows:
(2)
In formula (1) and (2),
With
Be respectively reference signal and sample signal quotient of amplitudes and the phase difference on frequency domain,
Be the refractive index of bactericide sample,
Be frequency,
Be the light velocity,
It is the bactericide sample thickness.
Described step 2): utilize principal component analysis method that frequency range is four kinds of bactericide absorption of sample spectrums of 0.3-1.6THz, 4*10 organizes data altogether, analyze, calculate each component, and with the first principal component PC1 and the second principal component PC2 as eigen vector, with radial base neural net four kinds of bactericide absorption spectras are discerned.
Described step 3): as eigen vector, utilize radial base neural net that four kinds of bactericide absorption spectras are discerned with the first principal component PC1 and the second principal component PC2.With in the neutral net identifying, every kind of bactericide absorption of sample is composed 6 groups of training that are used for radial base neural net in 10 eigenvectors, all the other 4 eigenvectors of every kind of bactericide absorption of sample spectrum are used for discerning.
The bactericide discrimination method that the present invention proposes is based on the method for terahertz time-domain spectroscopy detection technique, and the medicine pre-treatment is simple, processing ease, and calculation procedure is less.This method is a kind of class of solid bactericide fast and effectively agricultural chemicals sorting technique.
Description of drawings
Fig. 1 is a terahertz time-domain spectroscopy system and device index path;
Fig. 2 is the absorption spectrogram of good fortune magnesium iron;
Fig. 3 is the absorption spectrogram of thiophanate-methyl;
Fig. 4 is the absorption spectrogram of captan;
Fig. 5 is the absorption spectrogram of probenazole;
Fig. 6 is the two-dimentional scatter diagram of four kinds of bactericide first principal components and second principal component.
Embodiment
The femto-second laser that terahertz time-domain spectroscopy of the present invention (THz-TDS) system is made by U.S. Coherent company and the Model Z-1 type THz system of U.S. Zomega company development are formed, as shown in Figure 1.
The step of bactericide solid pesticide sorting technique is as follows:
1) utilize the terahertz time-domain system that purity is detected greater than the former medicine of 98% solid bactericide, obtain the time-domain signal of each sample, and calculate absorption spectra, bactericide original drug is ferbam, thiophanate-methyl, probenazole and captan;
2) utilize principal component analysis method that the absorption spectra of each sample is analyzed, extract feature, be used for the classification of each sample spectra;
3) utilize radial base neural net that the tera-hertz spectra of each bactericide sample is discerned.
Described step 1) is:
(1) dry in drying box respectively purity greater than 98% Coromate, methyl sulfur bacterium bactericide, probenazole bactericide and captan bactericide original drug and polyethylene powders, in mortar, grind, then four kinds of bactericide original drugs are mixed by 1:1 with polyethylene respectively, depress to the bactericide sample of thickness with tablet press machine at the pressure of 30MPa at 1.3mm, the quality of bactericide sample is 200mg, and diameter is 13mm;
(2) humidity less than 4% nitrogen environment in, utilize the terahertz time-domain system that empty specimen holder and the specimen holder that is placed with the bactericide sample are detected, each duplicate detection 10 times obtains empty specimen holder reference signal and the specimen holder sample signal that is placed with the bactericide sample respectively; Utilize Fourier transformation to convert reference signal and sample signal to frequency domain, calculate the absorption spectra of each bactericide sample with fresnel formula, formula is as follows:
(1)
In formula (1) and (2),
With
Be respectively reference signal and sample signal quotient of amplitudes and the phase difference on frequency domain,
Be the refractive index of bactericide sample,
Be frequency,
Be the light velocity,
It is the bactericide sample thickness.
Described step 2): utilize principal component analysis method that frequency range is four kinds of bactericide absorption of sample spectrums of 0.3-1.6THz, 4*10 organizes data altogether, analyze, calculate each component, and with the first principal component PC1 and the second principal component PC2 as eigen vector, with radial base neural net four kinds of bactericide absorption spectras are discerned.
Described step 3): as eigen vector, utilize radial base neural net that four kinds of bactericide absorption spectras are discerned with the first principal component PC1 and the second principal component PC2.With in the neutral net identifying, every kind of bactericide absorption of sample is composed 6 groups of training that are used for radial base neural net in 10 eigenvectors, all the other 4 eigenvectors of every kind of bactericide absorption of sample spectrum are used for discerning.
Embodiment
The step of bactericide solid pesticide method for identifying and classifying is as follows:
1) purity is dry in drying box respectively greater than 98% Coromate, methyl sulfur bacterium bactericide, probenazole bactericide and captan bactericide and polyethylene powders, in mortar, grind, then four kinds of bactericide original drugs are mixed by 1:1 with polyethylene respectively, depress to the bactericide sample of thickness with tablet press machine at the pressure of 30MPa at 1.3mm, the quality of bactericide sample is 200mg, and diameter is 13mm;
2) humidity less than 4% nitrogen environment under, utilize the terahertz time-domain spectroscopy system that empty specimen holder is detected, obtain time domain waveform, as the reference signal, ten groups of duplicate detection;
3) humidity less than 4% nitrogen environment under, utilize the terahertz time-domain spectroscopy system that the specimen holder that is placed with the bactericide sample is detected, obtain time domain waveform, as sample signal, ten groups of duplicate detection;
4) respectively reference signal and the sample signal of collecting carried out FFT, is converted to frequency-region signal,, calculate the absorption spectra of each sample according to fresnel formula:
(2)
In formula (1) and (2),
With
Be respectively reference signal and sample signal quotient of amplitudes and the phase difference on frequency domain,
Be the refractive index of bactericide sample,
Be frequency,
Be the light velocity,
It is the bactericide sample thickness;
5) utilize the absorption spectra of principal component analysis method to four kinds of bactericide samples, 4*10 organizes data altogether, analyze, extract feature, the eigen value contribution rate of first principal component and second principal component is respectively 87.93 and 10.47, preceding two construable variances of component have reached 98.4%, can explain most features of each sample.Be used for classification so choose the first principal component PC1 and the second principal component PC2 as eigen vector;
6) utilize radial base neural net that bactericide absorption of sample spectrum is discerned.In with the neutral net identifying, dispersion constant spread gets 0.1, under the less situation of sample size, reach higher recognition accuracy, and neuronal quantity will be more relatively, so here dispersion constant spread obtains less.Every kind of bactericide absorption of sample is composed in 10 eigenvectors 6 groups, and the 4*6 group is used for the training of radial base neural net altogether, all the other 4 eigenvectors of every kind of bactericide absorption of sample spectrum, and the 4*4 group is used for identification altogether.Recognition result sees Table 1.
Table 1
Sequence number | Medicine | M 1 | M 2 | M 3 | M 4 | |
1 | Good | 4 | 0 | 0 | 0 | 0% |
2 | Thiophanate- | 0 | 4 | 0 | 0 | 0% |
3 | | 0 | 0 | 4 | 0 | 0% |
4 | | 0 | 0 | 0 | 4 | 0% |
Claims (4)
1. bactericide solid pesticide sorting technique is characterized in that its step is as follows:
1) utilize the terahertz time-domain system that purity is detected greater than the former medicine of 98% solid bactericide, obtain the time-domain signal of each sample, and calculate absorption spectra, bactericide original drug is ferbam, thiophanate-methyl, probenazole and captan;
2) utilize principal component analysis method that the absorption spectra of each sample is analyzed, extract feature, be used for the classification of each sample spectra;
3) utilize radial base neural net that the tera-hertz spectra of each bactericide sample is discerned.
2. a kind of bactericide solid pesticide sorting technique according to claim 1 is characterized in that described step 1) is:
(1) dry in drying box respectively purity greater than 98% Coromate, methyl sulfur bacterium bactericide, probenazole bactericide and captan bactericide original drug and polyethylene powders, in mortar, grind, then four kinds of bactericide original drugs are mixed by 1:1 with polyethylene respectively, depress to the bactericide sample of thickness with tablet press machine at the pressure of 30MPa at 1.3mm, the quality of bactericide sample is 200mg, and diameter is 13mm;
(2) humidity less than 4% nitrogen environment in, utilize the terahertz time-domain system that empty specimen holder and the specimen holder that is placed with the bactericide sample are detected, each duplicate detection 10 times obtains empty specimen holder reference signal and the specimen holder sample signal that is placed with the bactericide sample respectively; Utilize Fourier transformation to convert reference signal and sample signal to frequency domain, calculate the absorption spectra of each bactericide sample with fresnel formula, formula is as follows:
(2)
3. a kind of bactericide solid pesticide sorting technique according to claim 1, it is characterized in that described step 2): utilize principal component analysis method that frequency range is four kinds of bactericide absorption of sample spectrums of 0.3-1.6THz, 4*10 organizes data altogether, analyze, calculate each component, and with the first principal component PC1 and the second principal component PC2 as eigen vector, with radial base neural net four kinds of bactericide absorption spectras are discerned.
4. a kind of bactericide solid pesticide sorting technique according to claim 1, it is characterized in that described step 3): with the first principal component PC1 and the second principal component PC2 as eigen vector, utilize radial base neural net that four kinds of bactericide absorption spectras are discerned, in with the neutral net identifying, every kind of bactericide absorption of sample is composed 6 groups of training that are used for radial base neural net in 10 eigenvectors, and all the other 4 eigenvectors of every kind of bactericide absorption of sample spectrum are used for identification.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201110178670A CN102283243A (en) | 2011-06-29 | 2011-06-29 | Solid bactericide pesticide classification method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201110178670A CN102283243A (en) | 2011-06-29 | 2011-06-29 | Solid bactericide pesticide classification method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN102283243A true CN102283243A (en) | 2011-12-21 |
Family
ID=45330223
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201110178670A Pending CN102283243A (en) | 2011-06-29 | 2011-06-29 | Solid bactericide pesticide classification method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN102283243A (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103424372A (en) * | 2013-08-26 | 2013-12-04 | 青岛科技大学 | Solid pesticide classification method based on decision tree algorithm |
CN104181122A (en) * | 2014-08-06 | 2014-12-03 | 首都师范大学 | Method for detecting content of pesticide in cereal by utilizing terahertz time-domain spectroscopy |
CN104237143A (en) * | 2013-06-08 | 2014-12-24 | 青岛科技大学 | Solid pesticide identification method based on terahertz spectroscopy |
CN104914068A (en) * | 2015-03-19 | 2015-09-16 | 哈尔滨商业大学 | Spectrum rapid detection method of trans-fatty acid content in grease |
CN109470647A (en) * | 2019-01-20 | 2019-03-15 | 南京林业大学 | A kind of measurement method of vapor Terahertz absorption spectra |
CN110751230A (en) * | 2019-10-30 | 2020-02-04 | 深圳市太赫兹科技创新研究院有限公司 | Substance classification method, substance classification device, terminal device and storage medium |
CN111272692A (en) * | 2019-12-11 | 2020-06-12 | 中国计量大学 | Method for detecting health product additive by using terahertz time-domain spectroscopy technology |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101876633A (en) * | 2009-11-13 | 2010-11-03 | 中国矿业大学 | Terahertz time domain spectroscopy-based textile fiber identification method |
-
2011
- 2011-06-29 CN CN201110178670A patent/CN102283243A/en active Pending
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101876633A (en) * | 2009-11-13 | 2010-11-03 | 中国矿业大学 | Terahertz time domain spectroscopy-based textile fiber identification method |
Non-Patent Citations (3)
Title |
---|
梁美彦: "利用人工神经网络识别毒品的THz光谱", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
王孝伟 等: "利用太赫兹时域光谱检测农产品中2种酰胺类农药残留", 《安徽农业科学》 * |
颜志刚 等: "戊唑醇和己唑醇的太赫兹(THz) 光谱研究", 《传感技术学报》 * |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104237143A (en) * | 2013-06-08 | 2014-12-24 | 青岛科技大学 | Solid pesticide identification method based on terahertz spectroscopy |
CN103424372A (en) * | 2013-08-26 | 2013-12-04 | 青岛科技大学 | Solid pesticide classification method based on decision tree algorithm |
CN104181122A (en) * | 2014-08-06 | 2014-12-03 | 首都师范大学 | Method for detecting content of pesticide in cereal by utilizing terahertz time-domain spectroscopy |
CN104914068A (en) * | 2015-03-19 | 2015-09-16 | 哈尔滨商业大学 | Spectrum rapid detection method of trans-fatty acid content in grease |
CN104914068B (en) * | 2015-03-19 | 2019-02-19 | 哈尔滨商业大学 | The spectrum rapid detection method of content of trans fatty acids in a kind of grease |
CN109470647A (en) * | 2019-01-20 | 2019-03-15 | 南京林业大学 | A kind of measurement method of vapor Terahertz absorption spectra |
CN110751230A (en) * | 2019-10-30 | 2020-02-04 | 深圳市太赫兹科技创新研究院有限公司 | Substance classification method, substance classification device, terminal device and storage medium |
CN111272692A (en) * | 2019-12-11 | 2020-06-12 | 中国计量大学 | Method for detecting health product additive by using terahertz time-domain spectroscopy technology |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102283243A (en) | Solid bactericide pesticide classification method | |
CN107677647B (en) | Method for identifying origin of traditional Chinese medicinal materials based on principal component analysis and BP neural network | |
CN102179375B (en) | Nondestructive detecting and screening method based on near-infrared for crop single-grain components | |
Cai et al. | Using FTIR spectra and pattern recognition for discrimination of tea varieties | |
Bhattacharya et al. | Preemptive identification of optimum fermentation time for black tea using electronic nose | |
CN104931470A (en) | Fluorescence hyperspectral technology-based pesticide residue detection device and method | |
Senthilkumar et al. | Detection of fungal infection and Ochratoxin A contamination in stored wheat using near-infrared hyperspectral imaging | |
Wetterich et al. | A comparative study on application of computer vision and fluorescence imaging spectroscopy for detection of Huanglongbing citrus disease in the USA and Brazil | |
CN106706546A (en) | Analysis method for artificial intelligence learning materials on basis of infrared and Raman spectrum data | |
Chen et al. | Fast detection of cumin and fennel using NIR spectroscopy combined with deep learning algorithms | |
Wang et al. | Method superior to traditional spectral identification: FT-NIR two-dimensional correlation spectroscopy combined with deep learning to identify the shelf life of fresh phlebopus portentosus | |
CN104020128A (en) | Method for rapidly identifying propolis source | |
CN103411906A (en) | Near infrared spectrum qualitative identification method of pearl powder and shell powder | |
Liu et al. | Discrimination of the fruits of Amomum tsao-ko according to geographical origin by 2DCOS image with RGB and Resnet image analysis techniques | |
Li et al. | Identification of pummelo cultivars by using Vis/NIR spectra and pattern recognition methods | |
CN104345045A (en) | Chemical pattern recognition and near infrared spectrum-based similar medicinal material identification method | |
Ma et al. | Efficient identification of Bachu mushroom by flourier transform infrared (FT-IR) spectroscopy coupled with PLS-GS-SVM | |
Ai et al. | Application of hyperspectral and deep learning in farmland soil microplastic detection | |
Yang et al. | Classification of wolfberry from different geographical origins by using electronic tongue and deep learning algorithm | |
Salman et al. | Detection of Fusarium oxysporum fungal isolates using ATR spectroscopy | |
Li et al. | Ultraviolet spectroscopy used to fingerprint five wild-grown edible mushrooms (Boletaceae) collected from Yunnan, China | |
Tee et al. | Discovering the Ganoderma boninense detection methods using machine learning: a review of manual, laboratory, and remote approaches | |
CN104374737A (en) | Near-infrared quantitative identification method | |
Agustika et al. | Classification of chili plant origin by using multilayer perceptron neural network | |
CN105486659A (en) | Construction method and application of corn seed variety authenticity identifying model |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20111221 |