CN108007917B - Method for establishing Raman spectrum measurement model of nitrogen content in rice plant by Hilbert method - Google Patents
Method for establishing Raman spectrum measurement model of nitrogen content in rice plant by Hilbert method Download PDFInfo
- Publication number
- CN108007917B CN108007917B CN201710156626.5A CN201710156626A CN108007917B CN 108007917 B CN108007917 B CN 108007917B CN 201710156626 A CN201710156626 A CN 201710156626A CN 108007917 B CN108007917 B CN 108007917B
- Authority
- CN
- China
- Prior art keywords
- raman spectrum
- spectrum
- nitrogen content
- rice
- raman
- 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.)
- Active
Links
Images
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
- G01N21/63—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light optically excited
- G01N21/65—Raman scattering
Landscapes
- Health & Medical Sciences (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating, Analyzing Materials By Fluorescence Or Luminescence (AREA)
Abstract
The invention discloses a method for establishing a Raman spectrum measurement model of nitrogen content in rice plants by a Hilbert method, and belongs to the technical field of measurement of trace elements in crops. According to the requirement of a measurement target, 4 functional modules are constructed: the device comprises a spectrum acquisition module, a spectrum preprocessing module, a spectrum data principal component analysis module and a pattern recognition module. The method mainly comprises the steps of collecting rice plant spectra by using a Raman spectrometer, denoising and baseline correcting Raman spectral data by using a wavelet decomposition method, extracting main components of the spectral data, carrying out Hilbert transform to obtain a frequency-wave number spectrum, obtaining nitrogen characteristic frequency by training and identifying a neural network, and establishing a least square measurement model of the characteristic frequency and the nitrogen content. The method for measuring the nitrogen content in the rice plant is not influenced by the contents of other trace elements in the rice, has the characteristics of micro-measurement, convenience and accuracy, and is suitable for batch operation.
Description
Technical Field
The invention relates to a method for establishing a Raman spectrum measurement model of nitrogen content in rice plants by a Hilbert method, belonging to the technical field of measurement of trace elements in crops.
Background
The traditional detection method of nitrogen level comprises a soil index method, a biochemical determination method and an empirical method. The soil index method is to supplement fertilizer to rice according to the nitrogen content in soil, namely a soil testing formula. The biochemical determination method is the main experimental method for detecting the nitrogen of the rice at present, and comprises a formaldehyde method, a Kjeldahl nitrogen determination method, a Dumas combustion nitrogen determination method and the like. The technical means or indexes of the empirical detection are mainly divided into three categories: visual inspection, remote spectrum sensing, and machine vision. The visual inspection method uses the abnormal color of the leaves as an index to diagnose the disease condition according to the production experience, and is a widely adopted detection means at present. The spectrum remote sensing method is characterized in that the spectrum of the rice field canopy is used as an analysis object, and the standard spectrum is compared to realize nitrogen deficiency detection. The machine vision method mainly uses a PC to distinguish the slight color difference between diseased rice and healthy rice plants from images, and is a digital extension of the visual method. In the method, only when rice has diseases, the spectrum remote sensing method and the machine vision method can be used for effective detection, namely, early measurement is difficult. Experience shows that after the rice shows the nitrogen deficiency symptom, the dosage of the additional fertilizer is at least doubled and obviously influences the growth of seedlings, so that the early detection of the nitrogen deficiency disease is very important. The soil index method is predictive, but the nitrogen in the rice seedlings is related to the nitrogen content in the soil, is influenced by illumination, temperature and humidity, soil pH value and the like, is a multi-factor constraint variable, and is an indirect measurement means with low measurement precision. The biochemical determination method is accurate and reliable, but has complex operation and low efficiency, can not carry out scale and rapid measurement and is mainly used in a control experiment. In the current rice nitrogen deficiency detection, although the visual detection method has large error, the operation is simple, and the method is the most widely applied detection means. There are two disadvantages to manual visual inspection: (1) the color of the rice plant needs to be observed according to experience by manual visual observation, and the judgment result is greatly influenced by subjective factors. When two or more nutrient elements are simultaneously deficient, the color of pathological changes is affected in a crossed manner, so that the visual inspection method cannot effectively judge the color. (2) Only serious nitrogen deficiency lesions can be observed by manual visual inspection. The change of rice plant expression is not obvious in early stage or light degree of pathological changes, but the change has serious influence on plant photosynthesis physiology and the like, and visual observation is difficult.
Disclosure of Invention
The invention aims to solve the problems in the prior art and research a Raman spectroscopy Hilbert frequency analysis method for the nitrogen level in rice plants, and the method can accurately, quickly and early measure the nitrogen level in the rice plants.
To achieve the above object, the present invention is achieved by the following steps:
(1) firstly, acquiring a Raman spectrum of a rice plant;
(2) then preprocessing the Raman spectrum data;
(3) then, data compression is carried out on the Raman spectrum signal to obtain a principal component;
(4) and finally, performing Hilbert transform on the main component, identifying to obtain the nitrogen characteristic frequency, and establishing a characteristic frequency and nitrogen content chemometric measurement model.
[ detailed description ] embodiments
The following provides a specific embodiment of the method for measuring the nitrogen content in rice plants based on the Hilbert frequency method of Raman spectrum.
In a dark room, at a constant temperature of 25 ℃, the laser wavelength is 785nm, the excitation power is 10mw, the rice leaves are pressed for 1 hour by a glass pressing sheet, and 0.5 × 0.5cm is cut from the middle part of the position 1cm away from the leaf tip2A leaf sample, namely placing the leaf sample on an objective table of a Raman spectrometer to obtain a Raman spectrum image of the sample; denoising and baseline drift removal processing are carried out on the Raman spectrum by using a wavelet decomposition method, and normalization is carried out; performing principal component regression analysis on the Raman spectrum signals, and obtaining principal component components through linear combination; and performing Hilbert transform on each principal component, performing convolution processing to obtain a standard frequency-wave number spectrum, obtaining nitrogen characteristic frequency through neural network training and recognition, and establishing a least square measurement model of the characteristic frequency and the nitrogen content.
The method has the beneficial effects that the scattering rule of the N-H antisymmetric and amide III deformation C-N bending vibration combined frequency can be accurately identified through the single-frequency component of the Raman spectrum, so that the spectrum frequency characteristic index of the nitrogen is established. The frequency characteristic index is different from the spectral wavelength characteristic index, cannot be influenced by characteristic peaks of other substances such as phosphorus, silicon, potassium and the like, has specificity and high identification precision, and is suitable for trace measurement of trace elements in plants.
Drawings
FIG. 1 is a functional block diagram of a method.
Detailed Description
The embodiments of the present invention will be described in detail below with reference to the accompanying drawings:
1. functional block diagram of method
The method is designed in a modularized mode so as to be convenient to adjust, reusable, easy to modify and easy to expand. According to the measurement target requirements, namely the Raman spectrum processing and modeling process of the rice plant leaves, 4 functional modules are constructed: the system comprises a spectrum acquisition module, a spectrum preprocessing module, a spectrum data principal component analysis module and a pattern recognition module, and is shown in figure 1.
The spectrum acquisition module realizes parameter adjustment, state setting, sample preparation, spectrum acquisition and data storage of the spectrum instrument; the spectrum preprocessing module provides a spectrum preprocessing algorithm, including denoising, baseline correction and normalization; the spectrum data principal component analysis module is used for compressing the spectrum data; the mode identification module screens out a nitrogen characteristic single frequency based on a time-frequency analysis method, and a nitrogen content measurement model is established by adopting a least square method.
2. Key technology
2.1 spectral pretreatment
2.1 wavelet decomposition denoising and de-baseline Drift
There are two types of noise in confocal micro-raman spectroscopy, electronic thermal motion noise from the instrument and external communication system interference. The existence of noise greatly affects the interpretation of the true information of the spectrum, so the signal noise reduction plays a significant role in spectrum analysis.
Signal reconstruction using sym1, sym2, sym3, sym4, sym5, sym6, sym7, sym8, db1, db2, db3, db4, db5, db6, db7, db8, db9, coif1, coif2, coif3, coif4, coif5 wavelet basis functions, respectively, using different threshold estimation methods hcursrc, Sqtwolog, rigrsurrc, Minimaxi, n (1-10) sub-decompositions, found that optimal denoising and de-baseline parameters were set for the rice leaf spectra in the present invention as follows: the basis function sym 8; threshold estimation method Hcurrurc; the number of decomposition layers was 5.
2.2 principal Components analysis of spectral data
And extracting characteristic values in the original spectrum data by using a principal component regression method and linearly combining again to realize the dimensionality reduction of the spectrum big data. Extraction of principal Components Using SPSS11.0, Default feature radical magnitude valuesThe quantity of extracted main components is 9, and the cumulative contribution rate reaches 97.31%. The accumulated contribution rate reflects that the principal component has strong expression capability on the original information, the colinearity among the spectrum data is effectively eliminated, and 9 principal components can effectively express the original spectrum information.
2.3 Hilbert transform principal component data and eigenfrequency screening
Will raman lightThe spectral wavelength signal is virtualized into a continuous time signal, andand performing convolution integral operation on the signals to obtain frequency components of Hilbert transform of each principal component. And identifying the association degree of each frequency component and the nitrogen through neural network training, setting a threshold value to screen out the characteristic frequency component, and establishing a partial least square model to measure the nitrogen content in the rice.
The neural network in the invention adopts a BP type structure, and the parameters are set as follows: the number of input layer neurons is determined by the characteristic frequency components, in this example, the input eigenvector is 9 × n; the output layer adopts 1 node, and 0 and 1 respectively represent that the nitrogen characteristic frequency component is false and true; the selection of the number of the hidden layer nodes has great influence on the performance of the network, the number of the nodes is too small, the nodes are easy to fall into local minimum values, the number of the hidden nodes is too large, a network fitting function is complex, the generalization capability of the network is poor, and the optimal number of the hidden layer nodes is determined according to a test result and is 8; the activation function of the neuron of the hidden layer selects a logsig () function, and the activation function of the output layer selects a pureline () function; the number of iterations is set to 1000, the network is trained 1 time every 10 steps, the target value is 0.01, the learning rate is 0.1, and the function of rainlm () is used as the training network.
Therefore, the method provided by the embodiment of the invention establishes the measurement model of the nitrogen content in the rice through the spectral characteristic frequency, is not influenced by the contents of other trace elements in the rice, has the characteristic of micro-measurement, and is suitable for batch operation. The invention is suitable for early detection of nitrogen deficiency diseases of crops such as rice and the like and can provide scientific basis for precision agriculture.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.
Claims (2)
1. A Raman spectrum measurement model method for establishing nitrogen content in rice plants by a Hilbert method is characterized by comprising the following steps of:
(1) firstly, acquiring a Raman spectrum of a rice plant;
(2) then decomposing the Raman spectrum by using a wavelet function, wherein the wavelet function basis function sym8 and a threshold estimation method Hcurrsurc have the decomposition layer number of 5, filtering wavelet high-frequency coefficients and realizing Raman spectrum denoising; filtering the wavelet low-frequency coefficient to remove the Raman spectrum baseline drift; normalizing the spectral data by using the spectral area to realize Raman spectral data preprocessing;
(3) performing principal component regression analysis on the Raman spectrum signals, and obtaining principal component components through linear combination;
(4) and performing Hilbert transform on each principal component, performing convolution processing to obtain a standard frequency-wave number spectrum, obtaining nitrogen characteristic frequency through neural network training and recognition, and establishing a least square measurement model of the characteristic frequency and the nitrogen content.
2. The method for establishing the Raman spectrum measurement model of the nitrogen content in the rice plant by the Hilbert method as claimed in claim 1, wherein the Raman spectrum of the rice plant is obtained by the following method:
in a dark room, under the constant temperature environment of 25 ℃;
the laser wavelength is 785nm, and the excitation power is 10 mw;
pressing the rice leaf for 1 hour by using a press glass sheet, and cutting 0.5 × 0.5.5 cm from the middle part of the position 1cm away from the leaf tip2Leaf samples;
and (4) placing the leaf sample on an objective table of a Raman spectrometer to obtain a Raman spectrum image of the sample.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710156626.5A CN108007917B (en) | 2017-03-16 | 2017-03-16 | Method for establishing Raman spectrum measurement model of nitrogen content in rice plant by Hilbert method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710156626.5A CN108007917B (en) | 2017-03-16 | 2017-03-16 | Method for establishing Raman spectrum measurement model of nitrogen content in rice plant by Hilbert method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108007917A CN108007917A (en) | 2018-05-08 |
CN108007917B true CN108007917B (en) | 2020-09-04 |
Family
ID=62048808
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710156626.5A Active CN108007917B (en) | 2017-03-16 | 2017-03-16 | Method for establishing Raman spectrum measurement model of nitrogen content in rice plant by Hilbert method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108007917B (en) |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111650129B (en) * | 2020-03-05 | 2022-07-01 | 广东省科学院广州地理研究所 | Litchi leaf calcium content prediction method and prediction device |
CN111413318B (en) * | 2020-04-30 | 2023-05-26 | 成都大象分形智能科技有限公司 | Serum detection system and method based on Raman spectrum |
CN113049500B (en) * | 2021-03-19 | 2022-12-06 | 杭州海康威视数字技术股份有限公司 | Water quality detection model training and water quality detection method, electronic equipment and storage medium |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102928396A (en) * | 2012-10-29 | 2013-02-13 | 浙江大学 | Urea isotopic abundance rapid detection method based on Raman spectrum |
KR20140038213A (en) * | 2012-09-20 | 2014-03-28 | 주식회사 스마테움 | Methods for assessing virus infections against a plant using raman spectroscopy and chemometrics |
CN104181151A (en) * | 2014-07-29 | 2014-12-03 | 蓝星化工新材料股份有限公司江西星火有机硅厂 | Method for online analyzing methyl chlorosilane by adopting online raman spectrometer to build model |
-
2017
- 2017-03-16 CN CN201710156626.5A patent/CN108007917B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20140038213A (en) * | 2012-09-20 | 2014-03-28 | 주식회사 스마테움 | Methods for assessing virus infections against a plant using raman spectroscopy and chemometrics |
CN102928396A (en) * | 2012-10-29 | 2013-02-13 | 浙江大学 | Urea isotopic abundance rapid detection method based on Raman spectrum |
CN104181151A (en) * | 2014-07-29 | 2014-12-03 | 蓝星化工新材料股份有限公司江西星火有机硅厂 | Method for online analyzing methyl chlorosilane by adopting online raman spectrometer to build model |
Non-Patent Citations (2)
Title |
---|
基于EMD和EEMD的自适应光谱预处理方法及其应用研究;赵肖宇;《中国博士学位论文全文数据库基础科学辑》;20160115(第1期);第5、16-18、53-60和73页 * |
水稻氮素营养水平与光谱特性的关系;周启发等;《浙江农业大学学报》;19931231(第S1期);第40-45页 * |
Also Published As
Publication number | Publication date |
---|---|
CN108007917A (en) | 2018-05-08 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102879353B (en) | The method of content of protein components near infrared detection peanut | |
CN102721651B (en) | Detection method and system of water content in plant leaf based on multispectral image | |
CN103940748B (en) | Based on the prediction of oranges and tangerines canopy nitrogen content and the visualization method of hyperspectral technique | |
CN108007917B (en) | Method for establishing Raman spectrum measurement model of nitrogen content in rice plant by Hilbert method | |
CN101915738A (en) | Method and device for rapidly detecting nutritional information of tea tree based on hyperspectral imaging technique | |
CN103278473B (en) | The mensuration of pipering and moisture and method for evaluating quality in white pepper | |
CN108519339B (en) | WT-L SSVR-based leaf cadmium content Vis-NIR spectral feature modeling method | |
CN103278503B (en) | Multi-sensor technology-based grape water stress diagnosis method and system therefor | |
Ning et al. | Rapid evaluation of soil fertility in tea plantation based on near-infrared spectroscopy | |
CN110363119A (en) | Based on wavelet transformation-random forests algorithm tobacco mildew method for quickly identifying | |
CN109142238B (en) | Cotton phosphorus nutrition rapid diagnosis method | |
Wang et al. | Intelligent detection of hard seeds of snap bean based on hyperspectral imaging | |
CN108007916B (en) | Confocal micro-Raman measurement model for establishing nitrogen content of rice plant by Hilbert-Huang method | |
Xiao et al. | Comparison of leaf chlorophyll content retrieval performance of citrus using FOD and CWT methods with field-based full-spectrum hyperspectral reflectance data | |
CN110596048A (en) | Method for quickly measuring potassium content in tobacco leaves by spectrum | |
Yang et al. | Classification of sugar beets based on hyperspectral and extreme learning machine methods | |
Liu et al. | Estimation of chlorophyll content in maize canopy using wavelet denoising and SVR method | |
CN102788796A (en) | Nutrient diagnosis device and nutrient diagnosis method for nitrogen of crops based on multi-information integration of high spectral images and fluorescent images | |
CN113267458A (en) | Method for establishing quantitative prediction model of soluble protein content of sweet potatoes | |
Zhou et al. | Hyperspectral imaging technology for detection of moisture content of tomato leaves | |
CN112782103B (en) | Method and system for monitoring early damage of cotton aphids on cotton seedling stage leaves | |
Chen et al. | Quality detection and variety classification of pecan seeds using hyperspectral imaging technology combined with machine learning | |
Lu et al. | Prediction performance optimization of different resolution and spectral band ranges for characterizing coco-peat substrate available nitrogen | |
Li et al. | Adulteration detection model of Tea Oil research based on FTIR back-propagation neural network | |
CN109324017B (en) | Method for improving near infrared spectrum analysis technology modeling spectrum quality |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |