CN101769867A - Nondestructive testing method for quality of compost products - Google Patents
Nondestructive testing method for quality of compost products Download PDFInfo
- Publication number
- CN101769867A CN101769867A CN201010033762A CN201010033762A CN101769867A CN 101769867 A CN101769867 A CN 101769867A CN 201010033762 A CN201010033762 A CN 201010033762A CN 201010033762 A CN201010033762 A CN 201010033762A CN 101769867 A CN101769867 A CN 101769867A
- Authority
- CN
- China
- Prior art keywords
- sample
- spectrum
- diffuse reflection
- quality
- reflection spectrum
- 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 relates to a nondestructive testing method for the quality of compost products, comprising the following steps: 1) utilizing a near infrared ray diffuse reflectance spectrum collection device to collect near infrared ray diffuse reflectance spectrums of representative samples of the compost products, and converting spectrum information into corresponding digital information; 2) determining standard content of modeling; 3) implementing the extraction and pretreatment for characteristic spectrums on the near infrared ray diffuse reflectance spectrums obtained in the step 1), to result in optimal spectrum data information in the near infrared ray diffuse reflectance spectrums; 4) regarding a spectrum data point combination, preferably selected in the step 3), as an input vector, and based on support vector machine regression method in combination with leave-one-out cross validation method, establishing a correction model for the contents of major technical indexes of the near infrared ray diffuse reflectance spectrums of the compost products to be tested, in combination with standard contents of samples measured in the step 2); 5) evaluating the model; and 6) collecting and inputting the near infrared ray diffuse reflectance spectrums of the compost products to be tested to the correction model established in the step 4), and computing the contents of major technical indexes predicating the samples to be tested. The nondestructive testing method can perform rapid and efficient testing on the quality of organic fertilizer products after high-temperature aerobic composting and is suitable for the rapid and nondestructive high-accuracy quantitative testing for the major technical indexes of organic fertilizer products after the high-temperature aerobic composting of agricultural and forest remainders.
Description
Technical field
The present invention relates to a kind of detection method of organic fertilizer product, particularly about a kind of lossless detection method of quality of compost products.
Background technology
Along with the intensivization development of livestock breeding industry, China's feces of livestock and poultry annual production presents ever-increasing trend.A large amount of feces of livestock and poultry very easily causes ecology and environmental pollution if deal with improperly.In recent years, the high temperature aerobic composting disposal route has become the important channel of the innoxious and recycling of a large amount of feces of livestock and poultry of China.Water percentage, organic matter, total nutrient content (nitrogen, phosphorus, potassium content) and potential of hydrogen etc. are to weigh the important technology index of quality of compost products.Traditional quality of compost products detection method complex operation step, time-consuming, require great effort and have certain contaminative.
Summary of the invention
At the problems referred to above, the purpose of this invention is to provide a kind of Non-Destructive Testing of quality of compost products fast and efficiently analytical approach.
For achieving the above object, the present invention takes following technical scheme: a kind of lossless detection method of quality of compost products, it may further comprise the steps: 1) utilize the near-infrared diffuse reflection spectrum harvester to gather the near-infrared diffuse reflection spectrum of composting production representative sample, and convert spectral information to corresponding numerical information; 2) according to agricultural industry relevant criterion method, determination step 1) in the key technical indexes content of selected representative sample as the standard content of modeling; 3) near-infrared diffuse reflection spectrum that step 1) is obtained carries out the extraction and the pre-service of characteristic spectrum, obtains the optimum spectroscopic data information in the near-infrared diffuse reflection spectrum; 4) the spectroscopic data point combination that step 3) is optimized is as input vector, utilize the method for the support vector machine Return Law in conjunction with the leaving-one method validation-cross, integrating step 2) the sample standard content that records in is set up a composting production near-infrared diffuse reflection spectrum the key technical indexes content calibration model; 5) performance of the composting production near-infrared diffuse reflection spectrum the key technical indexes content calibration model that step 4) is obtained is estimated; 6) gather the near-infrared diffuse reflection spectrum of composting production to be measured, input step 4) in the calibration model set up, calculate the key technical indexes content of prediction testing sample.
The parameter that composting production near-infrared diffuse reflection spectrum the key technical indexes content calibration model in the described step 4) comprises has: calibration set sample number n
c, checking collection sample number n
v, the chemical analysis value of i sample is actual value y
i, the near infrared measured value z of i sample
i, the mean value y of calibration set sample actual value
c, the mean value y of checking collection sample actual value
v, the mean value y of actual value, the mean value of near infrared predicted value
The major component factor that model uses is counted k.
In described step 3), that the spectrum pre-service is adopted is level and smooth, at least a in single order or second derivative, polynary scatter correction, variable standardization, the data centerization.
In described step 1), the near-infrared diffuse reflection spectrum harvester comprises Fourier transform near infrared spectrometer and integrating sphere diffuse reflection test sample device.
Each sample repeats filling scanning 3 times, gets the final spectrum of the averaged spectrum of 3 spectrum as counter sample.
Each scanning 32 times, the resolution of scanning is 8cm
-1, the spectral scan scope is 10000~4000cm
-1
The present invention is owing to take above technical scheme, it has the following advantages: 1, because the present invention extensively collects the composting production representative sample, use the near-infrared diffuse reflection spectrum of Fourier transform type near infrared spectrometer collected specimens, utilize genetic algorithm to carry out the preferred and spectroscopic data pre-service of characteristic spectrum to the spectroscopic data that obtains, the key technical indexes content that adopts the standard method working sample then is as standard content, utilize support vector machine to set up composting production the key technical indexes standard content near infrared spectrum calibration model as regression algorithm, therefore not only can extract composting production spectral signature information exactly, significantly reduce the quantity of information that participates in modeling, improve operation efficiency greatly, but also can be quick, harmless and accurately the composting production the key technical indexes is carried out detection by quantitative, the result shows, can realize the quick and efficient quantitative test of composting production the key technical indexes based on the quality of compost products near infrared spectrum speed cls analysis method of genetic algorithm and support vector machine.2, because the harvester of near-infrared diffuse reflection spectrum of the present invention comprises Fourier transform near infrared spectrometer and integrating sphere diffuse reflection test sample device, and scanning times is 32 times, resolution is 8cm
-1, the spectral scan scope is 10000~4000cm
-1Each sample repeats filling scanning 3 times, get the final spectrum of the averaged spectrum of 3 spectrum as counter sample, and the positive good utilisation of the inventive method near-infrared spectral analysis technology simple to operate, sample need not loaded down with trivial details pre-service, can finish the detection of a plurality of technical indicators of sample in short time simultaneously, analyte detection process is environment friendly and pollution-free, and can be used for the advantages such as online detection of sample.3,, it is hereby ensured removal interference farthest and extract the effective information of spectrum because preprocessing procedures of the present invention be at least a in level and smooth, single order or second derivative, polynary scatter correction, variable standardization, the data centerization.The present invention can detect fast and efficiently to the quality of organic fertilizer product behind the high temperature aerobic composting, be applicable to organic fertilizer product the key technical indexes behind the agricultural residue high temperature aerobic composting fast, harmless and high precision quantitatively detects.
Description of drawings
Fig. 1 is the near-infrared diffuse reflection spectrum figure of the composting production sample of one embodiment of the invention
Fig. 2 is composting production content of organic matter calibration set in the corresponding diagram 1 of the present invention and the scatter diagram that concerns of verifying collection chemical analysis value and spectroscopic assay value
Embodiment
Below in conjunction with drawings and Examples the present invention is described in detail.
The present invention includes following steps:
1) utilizes the near-infrared diffuse reflection spectrum harvester to gather the near-infrared diffuse reflection spectrum of composting production representative sample, and convert spectral information to corresponding numerical information.
These step concrete operations are: carry out the scanning of near-infrared diffuse reflection spectrum harvester background spectrum earlier, then Powdered or graininess composting production representative sample are placed sample cup, the near-infrared diffuse reflection spectrum of scanning representative sample converts spectral information to corresponding numerical information again.
Each representative sample repeats filling scanning 3 times, scans 32 times at every turn, gets the final near-infrared diffuse reflection spectrum of the averaged spectrum of 3 near-infrared diffuse reflection spectrums as counter sample, and the resolution of scanning is 8cm
-1, the spectral scan scope is 10000~4000cm
-1
The near-infrared diffuse reflection spectrum harvester comprises Fourier transform near infrared spectrometer, integrating sphere diffuse reflection test sample device and computing machine.The collection of near-infrared diffuse reflection spectrum and processing can utilize SPECTRUM ONE NTS signals collecting software and QUANT by Matlab7.0 platform and software package thereof
+Obtain with Unscrambler 9 data processing softwares.
The composting production representative sample is meant the effective sample that can contain each technical indicator content that China's agricultural industry relevant criterion requires as " organic fertilizer " and " biological organic fertilizer " or magnitude range and be evenly distributed.
2) according to agricultural industry relevant criterion method, determination step 1) in the key technical indexes content of selected representative sample as the standard content of modeling.
3) near-infrared diffuse reflection spectrum that step 1) is obtained carries out the extraction and the pre-service of characteristic spectrum, obtains the optimum spectroscopic data information in the near-infrared diffuse reflection spectrum.
In this step, the method for extracting characteristic spectrum is a genetic algorithm; Pre-service can be according to the situation of spectral quality and interference, selects at least a in level and smooth, single order or second derivative, polynary scatter correction, variable standardization, the data centerization etc.It is little fixed to extract characteristic spectrum and the pretreated sequencing of spectrum, is criterion with validation-cross standard deviation (SECV) minimum.Owing to comprised a lot of data points in the near-infrared diffuse reflection spectrum, extracting characteristic spectrum and carrying out the pretreated purpose of spectrum is in order to obtain the data point of each technical indicator quantitative test contribution rate maximum of quality of compost products to be made up, thereby makes precision of prediction higher.
4) the spectroscopic data point combination that step 3) is optimized is as input vector, utilize the method for the support vector machine Return Law and leaving-one method validation-cross, and integrating step 2) the sample standard content that records in is set up a composting production near-infrared diffuse reflection spectrum the key technical indexes content calibration model.The mutual use of the support vector machine Return Law and two kinds of methods of leaving-one method can avoid model " to owe match " or the generation of " over-fitting " phenomenon.
The parameter that composting production near-infrared diffuse reflection spectrum the key technical indexes content calibration model comprises has: calibration set sample number n
c, checking collection sample number n
v, the chemical analysis value of i sample is actual value y
i, the near infrared measured value z of i sample
i, the mean value y of calibration set sample actual value
c, the mean value y of checking collection sample actual value
v, the mean value y of actual value, the mean value of near infrared predicted value
The major component factor that model uses is counted k.
5) evaluation model: the performance of the composting production near-infrared diffuse reflection spectrum the key technical indexes content calibration model that step 4) is obtained is estimated.
The step of evaluation is as follows:
A, calculate decision calibration set coefficient of determination R
2With checking collection coefficient of determination r
2, represent the fitting degree that concerns between predicted value and the actual value respectively:
The predicted value of b, calculation correction collection and the deviation SEE between actual value, and the predicted value of checking collection and the deviation SEP between actual value:
Deviation when c, calculating leaving-one method validation-cross between predicted value and actual value, i.e. validation-cross standard deviation SECV:
D, the standard deviation that calculates checking collection sample actual value and the ratio of validation criteria difference, promptly relative analytical error RPD:
RPD=SD/SEP
E: chemistry assay value and near infrared measured value average poor, i.e. systematic error Bias:
The evaluation principle is: R
2And r
2Near 1, SEE, SEP, SECV and Bias are more little more, and the RPD value is big more, show that the precision of prediction of model is high more.
6) gather the near-infrared diffuse reflection spectrum of composting production to be measured, input step 4) in the calibration model set up, calculate the key technical indexes content of testing sample.
Be a specific embodiment below.
The inventive method is applied in the mensuration of the content of organic matter in the livestock excrement composting product.
1) gathers as shown in Figure 1, the near infrared diffuse spectrum of livestock excrement composting product representative sample.
Instrument: the SPECTRUM ONE of PE company Fourier transform near infrared spectrometer, integrating sphere diffuse reflection test sample device.
The condition of scanning: each representative sample repeats filling scanning 3 times, scans 32 times at every turn, gets the final near-infrared diffuse reflection spectrum of the averaged spectrum of 3 near-infrared diffuse reflection spectrums as counter sample, and the resolution of scanning is 8cm
-1, the spectral scan scope is 10000~4000cm
-1, every spectrum comprises 3001 data points.
2) utilize the agricultural industry relevant criterion, the mensuration modeling sees Table 1 with the key technical indexes content statistics of livestock excrement composting outturn sample, and unit is g kg
-1
Table 1
Sample number | Mean value | Standard deviation | Maximal value | Minimum value | |
All | ??120 | ??387.00 | ??142.11 | ??642.42 | ??100.37 |
Calibration set | ??90 | ??386.04 | ??142.94 | ??642.42 | ??100.37 |
The checking collection | ??30 | ??389.91 | ??141.94 | ??624.91 | ??126.49 |
3) sample spectra is carried out pre-service and extracted characteristic spectrum.
Carry out the sample spectra pre-service earlier, best practice is that first order derivative is handled; The genetic algorithm of carrying out characteristic spectrum again is preferred, and 500 data points in the preferred full spectrum are as the regression modeling input quantity.
4) the utilization support vector machine is set up content of organic matter regression model in the livestock excrement composting product.
Utilization gaussian kernel function (RBF nuclear) non-linear support vector machine homing method (v-SVM regression) is set up livestock excrement composting product content of organic matter regression model, and selected regression parameter and modeling result such as table 2 are listed.
Table 2
Number of data points | ??C | ??γ | ??R 2 | ??SEE | ??r 2 | ??SEP | ??RPD | ??Bias |
??500 | ??3.86E+07 | ??1.29E-02 | ??0.95 | ??30.93 | ??0.95 | ??31.86 | ??4.46 | ??1.69 |
5) to the evaluation of the model in the step 4).
As shown in Figure 2, the evaluating of model is as follows: checking collection coefficient of determination r
2, validation criteria difference SEP, relatively analytical error RPD and systematic error Bias.In Fig. 2, horizontal ordinate is represented the chemical analysis value, and ordinate is represented the spectroscopic assay value, and triangle is represented calibration set, and circle is represented forecast set.
As can be seen, the model of setting up has good predictive ability, and precision is higher, can well carry out actual condition and use.
Claims (7)
1. the lossless detection method of a quality of compost products, it may further comprise the steps:
1) utilizes the near-infrared diffuse reflection spectrum harvester to gather the near-infrared diffuse reflection spectrum of composting production representative sample, and convert spectral information to corresponding numerical information;
2) according to agricultural industry relevant criterion method, determination step 1) in the key technical indexes content of selected representative sample as the standard content of modeling;
3) near-infrared diffuse reflection spectrum that step 1) is obtained carries out the extraction and the pre-service of characteristic spectrum, obtains the optimum spectroscopic data information in the near-infrared diffuse reflection spectrum;
4) the spectroscopic data point combination that step 3) is optimized is as input vector, utilize the method for the support vector machine Return Law in conjunction with the leaving-one method validation-cross, integrating step 2) the sample standard content that records in is set up a composting production near-infrared diffuse reflection spectrum the key technical indexes content calibration model;
5) performance of the composting production near-infrared diffuse reflection spectrum the key technical indexes content calibration model that step 4) is obtained is estimated;
6) gather the near-infrared diffuse reflection spectrum of composting production to be measured, input step 4) in the calibration model set up, calculate the key technical indexes content of prediction testing sample.
2. a kind of lossless detection method of quality of compost products according to claim 1, it is characterized in that: the parameter that the composting production near-infrared diffuse reflection spectrum the key technical indexes content calibration model in the described step 4) comprises has: calibration set sample number n
c, checking collection sample number n
v, the chemical analysis value of i sample is actual value y
i, the near infrared measured value z of i sample
i, the mean value y of calibration set sample actual value
c, the mean value y of checking collection sample actual value
v, the mean value y of actual value, the mean value of near infrared predicted value
The major component factor that model uses is counted k.
3. a kind of lossless detection method of quality of compost products according to claim 1, it is characterized in that: in described step 3), that the spectrum pre-service is adopted is level and smooth, at least a in single order or second derivative, polynary scatter correction, variable standardization, the data centerization.
4. as the lossless detection method of a kind of quality of compost products as described in the claim 2, it is characterized in that: in described step 3), that the spectrum pre-service is adopted is level and smooth, at least a in single order or second derivative, polynary scatter correction, variable standardization, the data centerization.
5. as the lossless detection method of a kind of quality of compost products as described in claim 1 or 2 or 3 or 4, it is characterized in that: in described step 1), the near-infrared diffuse reflection spectrum harvester comprises Fourier transform near infrared spectrometer and integrating sphere diffuse reflection test sample device.
6. as the lossless detection method of a kind of quality of compost products as described in claim 1 or 2 or 3 or 4, it is characterized in that: each sample repeats filling scanning 3 times, gets the final spectrum of the averaged spectrum of 3 spectrum as counter sample.
7. as the lossless detection method of a kind of quality of compost products as described in the claim 6, it is characterized in that: scan 32 times, the resolution of scanning is 8cm at every turn
-1, the spectral scan scope is 10000~4000cm
-1
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201010033762A CN101769867A (en) | 2010-01-08 | 2010-01-08 | Nondestructive testing method for quality of compost products |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201010033762A CN101769867A (en) | 2010-01-08 | 2010-01-08 | Nondestructive testing method for quality of compost products |
Publications (1)
Publication Number | Publication Date |
---|---|
CN101769867A true CN101769867A (en) | 2010-07-07 |
Family
ID=42502860
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201010033762A Pending CN101769867A (en) | 2010-01-08 | 2010-01-08 | Nondestructive testing method for quality of compost products |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN101769867A (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101806738A (en) * | 2010-04-27 | 2010-08-18 | 南京农业大学 | Method for fast characterizing compost maturity |
CN102668902A (en) * | 2012-06-12 | 2012-09-19 | 杭州师范大学 | Method for evaluating inhibition ability of compost on soil-borne disease |
CN103575680A (en) * | 2013-11-22 | 2014-02-12 | 南京农业大学 | Spectroscopic method for evaluating quality indexes of organic fertilizer |
CN103837501A (en) * | 2014-03-12 | 2014-06-04 | 江苏绿威环保科技有限公司 | Analytical method of sludge water content |
CN106198447A (en) * | 2016-07-13 | 2016-12-07 | 中国科学院合肥物质科学研究院 | Chemical Mixed Fertilizer main component harmless quantitative detection method based on near-infrared spectrum technique |
CN108226090A (en) * | 2016-12-15 | 2018-06-29 | 中国农业机械化科学研究院 | A kind of method of component content detection model structure |
CN113804647A (en) * | 2021-09-18 | 2021-12-17 | 中国农业大学 | Online and offline detection method and system for liquid organic fertilizer |
-
2010
- 2010-01-08 CN CN201010033762A patent/CN101769867A/en active Pending
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101806738A (en) * | 2010-04-27 | 2010-08-18 | 南京农业大学 | Method for fast characterizing compost maturity |
CN101806738B (en) * | 2010-04-27 | 2011-05-25 | 南京农业大学 | Method for fast characterizing compost maturity |
CN102668902A (en) * | 2012-06-12 | 2012-09-19 | 杭州师范大学 | Method for evaluating inhibition ability of compost on soil-borne disease |
CN103575680A (en) * | 2013-11-22 | 2014-02-12 | 南京农业大学 | Spectroscopic method for evaluating quality indexes of organic fertilizer |
CN103837501A (en) * | 2014-03-12 | 2014-06-04 | 江苏绿威环保科技有限公司 | Analytical method of sludge water content |
CN103837501B (en) * | 2014-03-12 | 2016-05-25 | 江苏绿威环保科技有限公司 | A kind of analytical method of moisture percentage in sewage sludge |
CN106198447A (en) * | 2016-07-13 | 2016-12-07 | 中国科学院合肥物质科学研究院 | Chemical Mixed Fertilizer main component harmless quantitative detection method based on near-infrared spectrum technique |
CN108226090A (en) * | 2016-12-15 | 2018-06-29 | 中国农业机械化科学研究院 | A kind of method of component content detection model structure |
CN108226090B (en) * | 2016-12-15 | 2020-02-07 | 中国农业机械化科学研究院 | Method for constructing component content detection model |
CN113804647A (en) * | 2021-09-18 | 2021-12-17 | 中国农业大学 | Online and offline detection method and system for liquid organic fertilizer |
CN113804647B (en) * | 2021-09-18 | 2023-02-21 | 中国农业大学 | Online and offline detection method and system for liquid organic fertilizer |
CN113804647B8 (en) * | 2021-09-18 | 2023-05-02 | 中国农业大学 | Liquid organic fertilizer on-line and off-line detection method and system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN101762569A (en) | Non-destructive monitoring method of livestock excrement industrialized composting fermentation process | |
CN101769867A (en) | Nondestructive testing method for quality of compost products | |
Triolo et al. | Near Infrared Reflectance Spectroscopy (NIRS) for rapid determination of biochemical methane potential of plant biomass | |
CN103018195B (en) | Method for determination of PCTFE content in PBX explosive by near infrared spectrum | |
CN101210875A (en) | Damage-free measurement method for soil nutrient content based on near infrared spectra technology | |
CN101413885A (en) | Near-infrared spectrum method for rapidly quantifying honey quality | |
CN104596957A (en) | Estimation method for content of copper in soil on basis of visible-light near-infrared spectrum technology | |
CN102590129B (en) | Method for detecting content of amino acid in peanuts by near infrared method | |
CN102879340A (en) | Method for quickly detecting nutritional quality of root/stem crops on basis of near-infrared spectrum | |
CN102636454A (en) | Method for quickly measuring content of low carbon number fatty acid in edible oil by near infrared spectrum | |
CN101609042A (en) | Hand-held soil nutrient nondestructive measurement system based near infrared spectrum | |
CN105044050A (en) | Rapid quantitative analysis method for metallic elements in crop straw | |
CN104390927B (en) | The quick determination method of ash content in coal sample | |
CN106198447A (en) | Chemical Mixed Fertilizer main component harmless quantitative detection method based on near-infrared spectrum technique | |
CN105486655A (en) | Rapid detection method for organic matters in soil based on infrared spectroscopic intelligent identification model | |
CN107655851A (en) | A kind of method based on near-infrared spectrum technique quick detection lysine content | |
CN105021564A (en) | Method for determining content of ergosterol in tobacco based on near infrared spectroscopic analysis technology | |
CN105044024A (en) | Method for nondestructive testing of grape berries based on near infrared spectrum technology | |
CN201503392U (en) | Handheld soil nutrient nondestructive measurement device based on near infrared spectrum | |
CN105699319A (en) | Near infrared spectrum quick detection method for total moisture of coal based on gaussian process | |
CN104596979A (en) | Method for measuring cellulose of reconstituted tobacco by virtue of near infrared reflectance spectroscopy technique | |
CN104596975A (en) | Method for measuring lignin of reconstituted tobacco by paper-making process by virtue of near infrared reflectance spectroscopy technique | |
CN104266998A (en) | Near-infrared spectrum detection method for isocyanate group content in spandex prepolymer | |
CN103487398B (en) | A kind of analytical method of lysine fermentation liquor | |
CN104316492A (en) | Method for near-infrared spectrum measurement of protein content in potato tuber |
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 |
Open date: 20100707 |