CN104964950B - LIBS landwaste kind identification method based on multielement crest - Google Patents
LIBS landwaste kind identification method based on multielement crest Download PDFInfo
- Publication number
- CN104964950B CN104964950B CN201510314021.5A CN201510314021A CN104964950B CN 104964950 B CN104964950 B CN 104964950B CN 201510314021 A CN201510314021 A CN 201510314021A CN 104964950 B CN104964950 B CN 104964950B
- Authority
- CN
- China
- Prior art keywords
- landwaste
- spectrum
- neural network
- crest
- measured
- 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.)
- Expired - Fee Related
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/62—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light
- G01N21/71—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light thermally excited
- G01N21/718—Laser microanalysis, i.e. with formation of sample plasma
Landscapes
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Plasma & Fusion (AREA)
- Optics & Photonics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Engineering & Computer Science (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The present invention provides a kind of landwaste kind identification method, and it comprises the following steps:S1, spectrum of the N kind standard landwaste in the range of 200 950nm is measured using laser induced breakdown spectrograph;S2, linewidth parameters are carried out to N kind standard landwaste spectrum, find out the crest spectral line of the element of the forward predetermined number of wherein content;Input value of the crest spectral line group composite character spectrum of selected element as BP neural network, is trained to BP neural network, obtains BP neural network structure in S3, selection standard landwaste;S4, spectrum of the landwaste to be identified in the range of 200 950nm is measured using laser induced breakdown spectrograph;S5, linewidth parameters are carried out to landwaste spectrum to be identified, draw the spectral line with identical element in step S2;The spectral line crest of selected element is combined into input value of the characteristic spectrum as BP neural network in S6, selection landwaste to be measured, and landwaste to be measured is identified using obtained BP neural network, is identified result.
Description
Technical field
The present invention relates to landwaste type identification technical field, more particularly to a kind of induced with laser based on multielement crest is hit
Wear spectrum landwaste kind identification method.
Background technology
In recent years, with the rapid development of oil drilling new technology, the landwaste returned out by shaft bottom is very in small, broken bits, even
Into powdered so that traditional cutting description work becomes very difficult.
At present, LIBS, which is applied to rock type knowledge method for distinguishing, mainly has:First method is to utilize
LIBS combination partial least squares discriminant analysis, realizes the automatic identification to rock sample, its experiment taken
Model mainly has two kinds:First, full spectrum model, this kind of method recognition result is relatively high, but needs data to be processed to compare
It is more, while ambient noise has a great influence.Second method is peak strength and ratio model (characteristic model), this method data
Handle fairly simple, but recognition correct rate has declined.The third method is to choose the essential elements such as Si, Al, Ca, Fe, according to
These element spectral line of emission strength build characteristic variables, landwaste type identification is carried out in conjunction with neutral net.4th kind is to complete
Spectrum first carries out principal component analysis, recycles neutral net to carry out landwaste type identification.In these methods, although some recognition effects
It is relatively good, but the consuming time is compared in data processing, and then recognition effect is poor for other methods.
The content of the invention
In view of this, the present invention provide it is a kind of can take into account data-handling efficiency and recognition result accuracy based on more
The LIBS landwaste kind identification method of element crest.
A kind of landwaste kind identification method, it comprises the following steps:
S1, spectrum of the N kind standard landwaste in the range of 200-950nm is measured using laser induced breakdown spectrograph;
S2, linewidth parameters are carried out to N kind standard landwaste spectrum, find out the ripple of the element of the forward predetermined number of wherein content
Peak spectral line;
Input of the crest spectral line group composite character spectrum of selected element as BP neural network in S3, selection standard landwaste
Value, is trained to BP neural network, obtains BP neural network structure;
S4, spectrum of the landwaste to be identified in the range of 200-950nm is measured using laser induced breakdown spectrograph;
S5, linewidth parameters are carried out to landwaste spectrum to be identified, draw the spectral line with identical element in step S2;
The spectral line crest of selected element is combined into input of the characteristic spectrum as BP neural network in S6, selection landwaste to be measured
Value, is identified to landwaste to be measured using obtained BP neural network, is identified result.
Advantageous effects:The present invention is by by the crest spectral line group composite character of the higher element of content in standard landwaste
Input value of the spectrum as BP neural network, the input value as BP neural network are not that sample measures the full spectrum for carrying out spectrum,
It is not the singlet line intensity of element.Input value so using the characteristic spectral line that crest spectral line forms as BP neural network, its
Input data is much fewer than full spectral analysis method, but its recognition effect is also better than full spectrum analysis;This method input
Although data are more than the characteristic variable method that peak strength forms, its recognition effect will be far better than peak strength identification side
Method, although while input data add, increase and it is few, therefore time required for its identification do not have big increasing
Add, take into account data-handling efficiency and recognition result accuracy well.
Brief description of the drawings
Fig. 1 is the LIBS landwaste type identification based on multielement crest that embodiment of the present invention provides
Method flow diagram.
Embodiment
As shown in figure 1, a kind of landwaste kind identification method, it comprises the following steps:
S1, spectrum of the N kind standard landwaste in the range of 200-950nm is measured using laser induced breakdown spectrograph.Can
Selection of land, following 9 kinds of rock samples can be selected, standard landwaste is pressed into by tablet press machine.The type of rock sample, the place of production and its
The content of middle essential element such as following table.
S2, linewidth parameters are carried out to N kind standard landwaste spectrum, find out the ripple of the element of the forward predetermined number of wherein content
Peak spectral line.Alternatively, selected element includes Mn, Mg, Si, Fe, Al, Ti, Ca, Na, K in standard landwaste.
Input of the crest spectral line group composite character spectrum of selected element as BP neural network in S3, selection standard landwaste
Value, is trained to BP neural network, obtains BP neural network structure.
Alternatively, the step S3 includes:
Mn, Mg, Si, Fe, Al, Ti, Ca, Na, the centre wavelength of crest spectral line of K element take respectively:279.482、
285.213、288.158、301.898、309.271、365.350、393.367、588.995、766.490nm。
Wave-length coverage takes respectively:279.121-279.824、284.895-285.583、287.831-288.511、
301.595-302.232、308.923-309.594、365.056-365.668、393.067-393.647、588.382-
589.755th, 765.930-767.103nm, above-mentioned crest spectral line data is combined into characteristic spectrum.(1) number of element can be more
It can lack, 5-15 kind elements can;(2) centre wavelength of every kind of element can take different values, and wave-length coverage also can use difference
Value.The invention is not limited in this regard.
Every kind of rock sample takes (X+Y) to organize characteristic spectrum data, and wherein X, Y is positive integer, and common N × (X+Y) organizes characteristic light
Modal data;Take X group data before every kind of sample, the common N × X groups of N kind standard landwaste, the data training group as BP neural network;Often
Y group data after kind sample, the common N × Y groups of N kind standard landwaste, as the data detection group of BP neural network, enter to BP neural network
Row training, examine, obtain BP neural network structure.Alternatively, X, Y can be the integer more than 10, and X, Y value are bigger, are spent
Time it is longer, but effect is more preferable.
BP neural network can use Matlab softwares to realize, with spectral wavelength (or principal component factor) for input variable.
Spectroscopic data collection is divided into two groups during data analysis, one group of carry out BP neural network training, another group of carry out BP neural network knowledge
Not.By testing multigroup different BP neural network parameter value, the BP nerve nets containing at least one hidden layer are finally determined
Network structure, input layer number are the variables number of input, and node in hidden layer can be the integer more than input variable number, than
Can be such as 20, node in hidden layer is more, spends the time longer, but effect is more preferable, and output layer interstitial content is divided to two kinds of feelings
Condition:The first situation is 5 nodes, and 9 kinds of samples are divided into 5 classes;Second of situation is 9 nodes, realizes the mutual of 9 kinds of samples
Distinguish.
S4, spectrum of the landwaste to be identified in the range of 200-950nm is measured using laser induced breakdown spectrograph.
S5, linewidth parameters are carried out to landwaste spectrum to be identified, draw the spectral line with identical element in step S2.
The spectral line crest of selected element is combined into input of the characteristic spectrum as BP neural network in S6, selection landwaste to be measured
Value, is identified to landwaste to be measured using obtained BP neural network, is identified result.Selected element and step in step S6
It is identical in rapid S2.
Alternatively, landwaste to be measured is identified in step S6 including carrying out classification to landwaste to be measured and mutually distinguishing.
Alternatively, the crest spectral line combination of selected element is made up of the crest of each element in standard landwaste, according to each
The shapes and sizes of the crest of element determine the span of each peak wavelength.
The present invention implements have advantages below:
1st, BP god can be greatly reduced using factorial analysis or according to the progress feature extraction of the crest of part essential element
Input variable and shortening program runtime through network.
2nd, feature extraction is carried out to full spectrum according to the crest of part essential element, it is possible to reduce a large amount of interference letter in full spectrum
Breath, improve the discrimination of BP neural network.Sample is classified and discrimination that sample is mutually distinguished is not less than 98%.
3rd, BP neural network not only can be very good to classify to testing sample, can also well distinguish and treat test sample
Product.
To sum up, the present invention can retain most of useful information in spectrum, and reduce a large amount of interference informations, efficiently quick
The classification for realizing testing sample with mutually distinguish.
Professional further appreciates that, with reference to the unit of each example of the embodiments described herein description
And algorithm steps, can be realized with electronic hardware, computer software or the combination of the two, in order to clearly demonstrate hardware and
The interchangeability of software, the composition and step of each example are generally described according to feature in the above description.This
A little functions are performed with hardware or software mode actually, application-specific and design constraint depending on technical scheme.Specially
Industry technical staff can realize described function using distinct methods to each specific application, but this realization is not
The scope of the present invention should be exceeded.
Directly it can be held with reference to the step of method or algorithm that the embodiments described herein describes with hardware, processor
Capable software module, or the two combination are implemented.Software module can be placed in random access memory, internal memory, read-only storage,
Institute is public in electrically programmable ROM, electrically erasable ROM, register, hard disk, moveable magnetic disc, CD-ROM or technical field
In the storage medium for any other forms known.
It is understood that for the person of ordinary skill of the art, it can be conceived with the technique according to the invention and done
Go out other various corresponding changes and deformation, and all these changes and deformation should all belong to the protection model of the claims in the present invention
Enclose.
Claims (3)
1. a kind of landwaste kind identification method, it is characterised in that it comprises the following steps:
S1, spectrum of the N kind standard landwaste in the range of 200-950nm is measured using laser induced breakdown spectrograph;
S2, linewidth parameters are carried out to N kind standard landwaste spectrum, find out the crest spectrum of the element of the forward predetermined number of wherein content
Line;
Input value of the crest spectral line group composite character spectrum of selected element as BP neural network in S3, selection standard landwaste,
BP neural network is trained, obtains BP neural network structure;
S4, spectrum of the landwaste to be identified in the range of 200-950nm is measured using laser induced breakdown spectrograph;
S5, linewidth parameters are carried out to landwaste spectrum to be identified, draw the spectral line with identical element in step S2;
The spectral line crest of selected element is combined into input value of the characteristic spectrum as BP neural network in S6, selection landwaste to be measured,
Landwaste to be measured is identified using obtained BP neural network, is identified result;
Selected element includes Mn, Mg, Si, Fe, Al, Ti, Ca, Na, K in the standard landwaste;
The step S3 includes:Mn, Mg, Si, Fe, Al, Ti, Ca, Na, the centre wavelength of crest spectral line of K element take respectively:
279.482、285.213、288.158、301.898、309.271、365.350、393.367、588.995、766.490nm;Ripple
Long scope takes respectively:279.121-279.824、284.895-285.583、287.820-288.544、301.595-302.232、
308.923-309.594、365.056-365.668、393.067-393.647、588.382-589.755、765.930-
767.103nm, above-mentioned crest spectral line data is combined into characteristic spectrum;Every kind of rock sample takes X+Y group characteristic spectrum data, its
Middle X, Y are positive integer, and common N × (X+Y) organizes characteristic spectrum data;Take X group data before every kind of sample, the common N × X of N kind standard landwaste
Group, the data training group as BP neural network;Y groups data after every kind of sample, the common N × Y groups of N kind standard landwaste, as BP god
Data detection group through network, is trained to BP neural network, examines, and obtains BP neural network structure.
2. landwaste kind identification method as claimed in claim 1, it is characterised in that landwaste to be measured is identified in step S6
Including carrying out classification to landwaste to be measured and mutually distinguishing.
3. landwaste kind identification method as claimed in claim 2, it is characterised in that content is forward in standard landwaste default
The crest spectral line combination of several elements is made up of the crest of each selected element, according to the shape of the crest of each selected element and
Size determines the span of each peak wavelength.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510314021.5A CN104964950B (en) | 2015-06-10 | 2015-06-10 | LIBS landwaste kind identification method based on multielement crest |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510314021.5A CN104964950B (en) | 2015-06-10 | 2015-06-10 | LIBS landwaste kind identification method based on multielement crest |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104964950A CN104964950A (en) | 2015-10-07 |
CN104964950B true CN104964950B (en) | 2018-03-23 |
Family
ID=54218997
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510314021.5A Expired - Fee Related CN104964950B (en) | 2015-06-10 | 2015-06-10 | LIBS landwaste kind identification method based on multielement crest |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104964950B (en) |
Families Citing this family (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106707355A (en) * | 2015-11-12 | 2017-05-24 | 中石化石油工程技术服务有限公司 | Automatic lithology identification method of sedimentary rock |
CN105717074A (en) * | 2016-01-22 | 2016-06-29 | 浙江大学 | Method for identifying matcha powder and green tea powder by aid of laser-induced breakdown spectra on basis of characteristic wavelengths |
CN105938098A (en) * | 2016-07-07 | 2016-09-14 | 四川大学 | Rock soil ignition loss prediction method and system based on laser-induced breakdown spectroscopy |
CN105938099A (en) * | 2016-07-07 | 2016-09-14 | 四川大学 | Rock character judging method and system based on laser-induced breakdown spectroscopy |
CN106404748B (en) * | 2016-09-05 | 2019-03-05 | 华中科技大学 | A kind of multiline combination laser induced breakdown spectroscopy cereal crops Production area recognition method |
CN106525818B (en) * | 2016-10-13 | 2019-01-01 | 中国科学院上海技术物理研究所 | Based on multiple database support and the modified LIBS active agent analysis method of too many levels |
CN106841170B (en) * | 2016-12-05 | 2020-10-02 | 西北大学 | Coal ash type identification method based on wavelet neural network algorithm |
CN108844898A (en) * | 2018-03-13 | 2018-11-20 | 长江大学 | A kind of landwaste kind identification method and system |
CN108444953A (en) * | 2018-03-13 | 2018-08-24 | 长江大学 | Rice varieties method for quick identification based on laser induced breakdown spectroscopy differential signal |
CN108596021A (en) * | 2018-03-13 | 2018-09-28 | 长江大学 | A kind of landwaste kind identification method and system based on laser induced breakdown spectroscopy differential signal |
CN108573105A (en) * | 2018-04-23 | 2018-09-25 | 浙江科技学院 | The method for building up of soil heavy metal content detection model based on depth confidence network |
CN108596085A (en) * | 2018-04-23 | 2018-09-28 | 浙江科技学院 | The method for building up of soil heavy metal content detection model based on convolutional neural networks |
CN109696425A (en) * | 2019-01-25 | 2019-04-30 | 长江大学 | A kind of landwaste kind identification method and system based on laser induced breakdown spectroscopy |
CN109975273A (en) * | 2019-03-07 | 2019-07-05 | 四川大学 | A kind of petrographic classification method based on laser induced breakdown spectroscopy |
CN111239103A (en) * | 2020-01-21 | 2020-06-05 | 上海海关工业品与原材料检测技术中心 | Method for identifying iron ore production country and brand |
CN112782151B (en) * | 2021-02-22 | 2023-01-13 | 湖北工程学院 | Data processing method for improving classification accuracy of laser-induced breakdown spectroscopy |
CN113177919B (en) * | 2021-04-28 | 2022-08-05 | 成都艾立本科技有限公司 | Lithology classification and principal component element content detection method combining LIBS and deep learning |
CN114118309B (en) * | 2022-01-28 | 2022-04-01 | 津海威视技术(天津)有限公司 | Sample classification and identification method based on convolutional neural network |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101915753A (en) * | 2010-07-30 | 2010-12-15 | 浙江师范大学 | Genetic Neural NetworkQuantitative analysis method for laser induced breakdown spectroscopy based on gGenetic Neural Networkgenetic neural network |
CN102830096A (en) * | 2012-08-29 | 2012-12-19 | 国电燃料有限公司 | Method for measuring element concentration and correcting error based on artificial neural network |
-
2015
- 2015-06-10 CN CN201510314021.5A patent/CN104964950B/en not_active Expired - Fee Related
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101915753A (en) * | 2010-07-30 | 2010-12-15 | 浙江师范大学 | Genetic Neural NetworkQuantitative analysis method for laser induced breakdown spectroscopy based on gGenetic Neural Networkgenetic neural network |
CN102830096A (en) * | 2012-08-29 | 2012-12-19 | 国电燃料有限公司 | Method for measuring element concentration and correcting error based on artificial neural network |
Non-Patent Citations (3)
Title |
---|
Accurate Identification of geological samples using artificial neural network processing of laser induced breakdown;Siu L L, Alexander K.;《J Anal At Spectrom》;20111231;第2419-2427页 * |
基于主成分分析和人工神经网络的激光诱导击穿光谱塑料分类识别方法研究;王茜蒨等;《光谱学与光谱分析》;20121231;第32卷(第12期);第3179-3180页第1节 * |
激光诱导击穿光谱在地质录井岩性快速识别中的应用;陈兴龙等;《中国激光》;20131231;第40卷(第12期);第244页第1节,第244-245页第2.1-2.2节,第245页倒数第1-2段 * |
Also Published As
Publication number | Publication date |
---|---|
CN104964950A (en) | 2015-10-07 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104964950B (en) | LIBS landwaste kind identification method based on multielement crest | |
CN105181678B (en) | Rice varieties discrimination method based on LIBS | |
CN108021938A (en) | A kind of Cold-strip Steel Surface defect online detection method and detecting system | |
CN103578098B (en) | Method and device for extracting commodity body in commodity picture | |
CN107655850A (en) | Non-linear modeling method and system based near infrared spectrum | |
Kenchington et al. | Rangeomorph classification schemes and intra-specific variation: are all characters created equal? | |
Özkan et al. | Identification of wheat kernels by fusion of RGB, SWIR, and VNIR samples | |
CN107657285A (en) | Hyperspectral image classification method based on Three dimensional convolution neutral net | |
CN103577993B (en) | Color choosing method and device | |
CN105548066A (en) | Method and system for distinguishing colloid types | |
CN109859204A (en) | Convolutional neural networks Model Checking and device | |
CN107229910A (en) | A kind of remote sensing images icing lake detection method and its system | |
CN103487410B (en) | Based on the Laser-induced Breakdown Spectroscopy Variable Selection of Models Sets cluster analysis | |
CN106680693A (en) | Batch test method for semiconductor chips based on grayscale map of fault data | |
CN116167964A (en) | Tumor classification method and system based on tumor hyperspectral image | |
CN105447513B (en) | A kind of file ink data realizes the method and system of automatic contrast | |
CN117611893A (en) | Remote sensing extraction method for soybean planting area based on redundancy-reducing feature optimization strategy | |
CN113077271A (en) | Enterprise credit rating method and device based on BP neural network | |
Huang et al. | Fast location and segmentation of high‐throughput damaged soybean seeds with invertible neural networks | |
JP2006521785A5 (en) | ||
CN113138181B (en) | Method for grading quality of fen-flavor wine base | |
CN106295667A (en) | A kind of method and application thereof selecting optimum spectrum based on genetic algorithm | |
CN103092854B (en) | Music data sorting method | |
US20120110531A1 (en) | Defect and yield prediction for segments of an integrated circuit | |
Stąpor | Better alternatives for stepwise discriminant analysis |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180323 Termination date: 20200610 |