CN109632699B - Bark phloem-based near-infrared detection model establishment method for navel orange yellow dragon disease - Google Patents
Bark phloem-based near-infrared detection model establishment method for navel orange yellow dragon disease Download PDFInfo
- Publication number
- CN109632699B CN109632699B CN201811611126.7A CN201811611126A CN109632699B CN 109632699 B CN109632699 B CN 109632699B CN 201811611126 A CN201811611126 A CN 201811611126A CN 109632699 B CN109632699 B CN 109632699B
- Authority
- CN
- China
- Prior art keywords
- bark
- navel orange
- spectrum
- sample
- phloem
- 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
- 201000010099 disease Diseases 0.000 title claims abstract description 42
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 title claims abstract description 42
- 240000002319 Citrus sinensis Species 0.000 title claims abstract description 38
- 235000005976 Citrus sinensis Nutrition 0.000 title claims abstract description 38
- 238000000034 method Methods 0.000 title claims abstract description 36
- 238000001514 detection method Methods 0.000 title claims abstract description 30
- 238000001228 spectrum Methods 0.000 claims abstract description 58
- 238000012937 correction Methods 0.000 claims abstract description 26
- 230000003595 spectral effect Effects 0.000 claims abstract description 22
- 235000013399 edible fruits Nutrition 0.000 claims abstract description 17
- 238000010606 normalization Methods 0.000 claims abstract description 14
- 238000007781 pre-processing Methods 0.000 claims abstract description 13
- 238000002835 absorbance Methods 0.000 claims description 9
- 230000000903 blocking effect Effects 0.000 claims description 5
- 238000004458 analytical method Methods 0.000 claims description 4
- 238000005070 sampling Methods 0.000 abstract description 15
- 238000010238 partial least squares regression Methods 0.000 abstract description 7
- 244000052616 bacterial pathogen Species 0.000 abstract description 4
- 238000013499 data model Methods 0.000 abstract description 4
- 241000207199 Citrus Species 0.000 description 17
- 235000020971 citrus fruits Nutrition 0.000 description 17
- 238000012628 principal component regression Methods 0.000 description 7
- 241001478315 Candidatus Liberibacter asiaticus Species 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 3
- 238000002329 infrared spectrum Methods 0.000 description 3
- 229910000530 Gallium indium arsenide Inorganic materials 0.000 description 2
- 230000015572 biosynthetic process Effects 0.000 description 2
- 238000005034 decoration Methods 0.000 description 2
- 238000003745 diagnosis Methods 0.000 description 2
- 239000003814 drug Substances 0.000 description 2
- 208000015181 infectious disease Diseases 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 208000024891 symptom Diseases 0.000 description 2
- 238000003786 synthesis reaction Methods 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 241000894006 Bacteria Species 0.000 description 1
- 241000196324 Embryophyta Species 0.000 description 1
- 240000001140 Mimosa pudica Species 0.000 description 1
- 235000016462 Mimosa pudica Nutrition 0.000 description 1
- 206010028980 Neoplasm Diseases 0.000 description 1
- 201000011510 cancer Diseases 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 229940079593 drug Drugs 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000003771 laboratory diagnosis Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 244000052769 pathogen Species 0.000 description 1
- 230000003950 pathogenic mechanism Effects 0.000 description 1
- 238000002203 pretreatment Methods 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 238000007789 sealing Methods 0.000 description 1
- 241000894007 species Species 0.000 description 1
- 238000004611 spectroscopical analysis Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
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
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
- G01N21/35—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light
- G01N21/359—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light using near infrared light
Landscapes
- Physics & Mathematics (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Health & Medical Sciences (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 Or Analysing Materials By Optical Means (AREA)
Abstract
The invention provides a method for establishing a navel orange yellow dragon disease near-infrared detection model, and relates to the technical field of data model establishment. The establishing method comprises the following steps: 1) taking navel orange trees infected with yellow dragon germs and navel orange trees not infected with yellow dragon germs as standard fruit trees respectively, and taking barks on side branches of the fruit trees as samples respectively; 2) respectively taking 5 points on the phloem surface of the sample to perform spectrum scanning, and collecting spectra; 3) after randomly dividing the collected spectral data into a correction set and a prediction set, preprocessing the data by using a normalization method, and establishing a partial least square regression model. The near-infrared prediction model of the navel orange yellow dragon disease based on the bark phloem is established by taking the bark phloem as a sample and adopting a Partial Least Squares Regression (PLSR), and the prediction accuracy and the data fitting degree of the model can keep good levels by comparing with models obtained from other sampling parts.
Description
Technical Field
The invention belongs to the technical field of data model establishment, and particularly relates to a near-infrared detection model establishment method for navel orange yellow dragon disease based on bark phloem.
Background
The citrus greening disease is the most dangerous disease faced by the citrus industry in the world, the pathogenic mechanism is not clear, no effective drug can treat the citrus greening disease at present, the citrus greening disease is commonly called citrus cancer, and the citrus greening disease faces the threat of the citrus greening disease in the main countries of the americas, brazil, australia and the like internationally. Citrus huanglongbing in China is spread throughout the whole world of citrus producing areas in Fujian, Guangdong, Guangxi and Hainan, is reported in more than ten provinces and areas in Zhejiang, Jiangxi, Hunan and the like, and spreads north year by year. Because no effective treatment medicine exists, the prevention and control of the disease are mainly performed, and the timely discovery of diseased trees and the adoption of measures such as cutting trees are the key points for preventing and treating the Huanglong disease, but the current diagnosis methods have certain limitations: the laboratory diagnosis method has low practicability due to long detection period, high cost and the like; the field symptom observation method is greatly influenced by personal experience and subjectivity, and particularly under the condition of no specific symptom, the result is unreliable.
The application of the near infrared spectrum technology to the field detection of citrus greening disease has been reported, Sindhuja Sankaran, W.Windham, Liuyande and the like all prove the feasibility of the near infrared detection technology in the diagnosis of citrus greening disease, and researches by people such as Ma 281111and sensitive plants show that rapid and nondestructive detection can be realized in the field by collecting spectral data of citrus leaves and modeling. The citrus huanglongbing usually occurs in phloem, and germs can be detected in leaves and barks all the year round, but the current research on the near-infrared detection of the citrus huanglongbing is only limited to take the leaves as samples.
The distribution rule of the Huanglongbing pathogenic bacteria in the citrus tree is not clear, and the integration of the spectral information of multiple parts is favorable for the positive detection rate of the Huanglongbing pathogenic bacteria. Johnson et al found higher levels of root pathogens, while Ding et al found the highest petiole content and the lowest root content, with the opposite conclusions. Moreover, the content of the yellow dragon disease bacteria in different tissues of the same tree is obviously different, for example, the pathogenic bacteria content in the plant body is found to be different from 308 cells/mu g DNA (pollen) to 109842 cells/mu g DNA (seed coat tissue) by the ocean of the Roebinella and the like. At present, a data model establishing method which can give consideration to both precision and fitting degree does not exist.
Disclosure of Invention
In view of this, the invention aims to provide a near-infrared detection model establishing method for citrus greening disease based on bark phloem, which can improve the precision and data fitting degree of a data model and further optimize the application of a near-infrared technology in the aspect of citrus greening disease detection.
In order to achieve the above object, the present invention provides the following technical solutions:
the invention provides a near-infrared detection model building method for navel orange yellow dragon disease based on bark phloem, which comprises the following steps:
1) respectively taking diseased navel orange trees and non-diseased navel orange trees as standard fruit trees, and respectively stripping barks on side branches of the fruit trees as samples, wherein the width of each sample is 0.4-0.6 cm;
2) respectively taking 5 points on the phloem surface of the sample to perform spectrum scanning, and collecting spectra; the resolution of the scan is 8cm-1The scanning speed is 5 spectra/second;
3) after randomly dividing the collected spectral data into a correction set and a prediction set, preprocessing the data by using a normalization method, and establishing a partial least square regression model.
Preferably, during the spectrum scanning in step 2), the sample is covered above a lens of the detector, and a light blocking cover is covered above the sample.
Preferably, the spectrum range of the collected spectrum in the step 2) is 12500-350 cm-1Spectral resolution of 0.25cm-1Spectral accuracy of 0.05cm-1Spectral accuracy of 0.1cm-1。
Preferably, before the data preprocessing in step 3), the method further comprises analyzing the collected original spectrum, wherein the analyzing comprises calculating an absorbance average of the spectrum by using the collected original spectrumThen, calculating D-Bark, and obtaining a spectrogram by using the D-Bark:
the invention provides a near-infrared detection model building method for navel orange yellow dragon disease based on bark phloem, which takes bark as a sample, and adopts a partial least squares regression method (PLSR) to build a navel orange yellow dragon disease prediction model based on the bark phloem after optimizing a data preprocessing method. In the embodiment of the invention, three sample schemes of leaf, bark and synthesis (leaf + bark) are designed, and the Root Mean Square Error (RMSEP) of the prediction set of the model is 10-5Magnitude, and root mean square error of the blade (RMSEPL, 1.6909 x 10)-5) < root mean square error of bark (RMSEPB,1.8890 x 10)-5) < Total root mean square error (RMSEPC,2.5676 x 10)-5) Coefficient (r) is determined for prediction set2) All above 0.9, and leaf blades (rL)20.9396) < bark (rB)20.9415) < Synthesis (rC20.9603), which shows that the models built by the three sampling schemes have good accuracy and prediction capability, while the model obtained by the leaf sampling scheme has the highest accuracy but the weakest prediction capability, while the model obtained by the comprehensive sampling scheme has the strongest prediction capability but the lowest model accuracy, and only the accuracy of the model obtained by the bark phloem sampling scheme (RMSEPB is 1.8890 x 10)-5) Prediction ability (rB)20.9415) can be kept at a good level.
Drawings
FIG. 1 is a standardized debarker and resulting bark sample;
FIG. 2 is a diagram of a spectral scanning spot distribution according to an embodiment of the present invention;
FIG. 3 is a graph of the original spectrum of bark in an embodiment of the present invention;
FIG. 4 is a raw spectrum of a blade in an embodiment of the present invention;
fig. 5 is an enlarged view of the spectral difference in the example of the present invention.
Detailed Description
The invention provides a near-infrared detection model building method for navel orange yellow dragon disease based on bark phloem, which comprises the following steps: 1) respectively taking diseased navel orange trees and non-diseased navel orange trees as standard fruit trees, and respectively stripping barks on side branches of the fruit trees as samples, wherein the width of each sample is 0.4-0.6 cm;
2) respectively taking 5 points on the phloem surface of the sample to perform spectrum scanning, and collecting spectra; the resolution of the scan is 8cm-1The scanning speed is 5 spectra/second;
3) after randomly dividing the collected spectral data into a correction set and a prediction set, preprocessing the data by using a normalization method, and establishing a partial least square regression model.
According to the establishing method, a diseased navel orange tree and an uninfected navel orange tree are respectively used as standard fruit trees, barks on side branches of the fruit trees are respectively stripped to be used as samples, and the width of each sample is 0.4-0.6 cm. According to the standard fruit tree disclosed by the invention, the infection condition of the yellow dragon disease is preferably confirmed by a real-time fluorescent PCR standard detection method, and the tree age, the tree species and the planting environment are consistent. The PCR standard detection method is not particularly limited in the present invention, and the detection is preferably performed by GB/T28062-2011 method. The invention is preferably carried out on each lateral branch of the fruit tree in four directions when stripping the sample, so as to ensure that the bark sampling does not influence the growth of the fruit tree, and the sampling is shown in figures 1-2. The method of collecting the sample according to the present invention is not particularly limited, and a standardized bark peeler is preferably used, and more preferably, an adjustable novel bark girdling device is used. The samples collected according to the present invention are preferably as shown in FIGS. 1-3, each having a width of 0.5cm and an average length of 11.34cm, with a partial enlargement as shown in FIGS. 1-4. In the embodiment of the invention, 291 parts of uninfected bark samples and 226 parts of infected bark samples on a New-Holer navel orange tree planted in 2011 are collected in a navel orange planting base of the national navel orange engineering and technology research center; in addition, in order to facilitate comparison, the leaves are collected as comparison samples, including the leaves on the fruit trees without diseases and the leaves on the fruit trees with diseases.
Respectively taking 5 points on the phloem surface of the sample to perform spectrum scanning, and collecting spectra; the resolution of the scan is 8cm-1The speed of the scan is 5 spectra/sec. The invention uniformly takes 5 points on the surface of the bark phloem for spectral scanning, and the resolution of the scanning is 8cm-1The speed of the scan is 5 spectra/sec. In the present invention, when performing the spectrum scanning, it is preferable that the sample is covered over a lens of the detector, and a light blocking cover is covered over the sample. The light blocking cover can avoid the influence of an external light source. The spectral range of the collected spectrum is preferably 12500-350 cm-1Spectral resolution of 0.25cm-1Spectral accuracy of 0.05cm-1Spectral accuracy of 0.1cm-1. In the present invention, preferably, each point is collected three times during the collection of the spectrum, and the corresponding average value is used as the spectrum data for analysis. The device for acquiring the spectrum preferably consists of a desktop computer and an IRTracer 100 NIR spectrometer, wherein the spectrum analyzer is a Fourier near infrared spectrum system, and the detector is an InGaAs detector. The spectral scanning points are preferably as shown in figures 2-3, five points are uniformly taken along the phloem surface of the bark, so that the bark is evenly divided into 4 equal parts, and the influence of a sampling part on an experimental result is avoided. In the embodiment of the invention, spectrum collection is also carried out on the blade of the comparison sample, and during the collection, 5-point spectrum scanning is also adopted, and the scanning points are shown in figures 2-1 and 2-2: three points are taken along the main pulse, two ends (A, C) and a middle point (B), and a point (D, E) is taken on the left side and the right side of the leaf surface respectively, so that the straight line where the BDE three points are located is vertical to the main pulse.
The invention randomly divides the collected spectral data into a correction set and a prediction set, preprocesses the data by a normalization method, and establishes a partial least square regression model by using Unscamblebler 10 software. The invention preferably further comprises, before said data preprocessing, the step of acquiring the raw spectraPerforming an analysis comprising averaging absorbance of the spectrum using the collected raw spectrumThen, after calculating D-Leaf, D-Bark, D-Hlb, D-Nhlb and D, forming a spectrum for analysis and comparison:
D=D-Leaf+D-Bark。
the invention obtains the utilization spectrumThe method (2) is not particularly limited, and may be a method conventionally used in the art. In the embodiment of the invention, the trends of the spectral peak shape and the peak position acquired by using the bark phloem surface are similar, and the peak value appears at 6875cm-1Left and right and 5150cm-1On the left and right, the obtained bark phloem spectrum has certain characteristics. Moreover, the spectral data of the infected sample and the spectrum data of the uninfected sample have no obvious difference and cannot be distinguished by naked eyes; however, after amplifying the spectrogram of D-Bark by 100 times, the absorbance value of the affected Bark can be found to be larger than that of the uninfected Bark.
In the embodiment of the invention, 345 correction sets (151 samples with yellow dragon disease and 194 normal samples without disease) of bark phloem samples and 172 prediction sets (75 samples with yellow dragon disease and 97 normal samples without disease) of bark phloem samples are randomly divided into a correction set and a prediction set.
The invention preprocesses the spectrum of the sample, and the data preprocessing preferably comprises: and (4) a normalization method. The normalization method can eliminate the influence of instruments, environmental conditions, sample background and other factors, and reduce phenomena of spectrogram baseline translation, drift, high-frequency random noise and the like as far as possible. The invention obtains a partial least square regression model by using data preprocessed by a normalization method, the partial least square regression model is preferably established by using Unscamblebler 10 software, and the established model correction set determines a coefficient R20.9390, the correction set root mean square error RMSEC is 2.0397 x 10-5Coefficient r is determined for prediction set20.9414, the prediction set root mean square error RMSEP is 1.8890 x 10-5。
The method for establishing the near-infrared detection model of navel orange yellow dragon disease based on bark phloem according to the present invention is described in detail with reference to the following examples, but they should not be construed as limiting the scope of the present invention.
Example 1
Step 1: sampling method for different navel orange detection parts
A. Bark sampling: with the standardized bark peeler as shown in fig. 1-1, the side branches of the fruit tree in four directions as shown in fig. 1-2 are used as sampling sites.
B. Leaf sampling: taking the canopy leaves in the east, west, south and north directions, numbering all samples, sealing and storing at 4 ℃ by using a sampling bag.
Step 2: selection of test materials and data statistics
The leaves and barks of navel oranges for experiments are sampled in a greenhouse and a navel orange planting base of the national navel orange engineering technology research center, all trees are the New Youer navel oranges planted in 2011, and the infection condition of the Huanglongbing disease is confirmed by a PCR standard detection method. The sampling cases are shown in table 1: 365 parts of infected leaf sample, 226 parts of infected tree phloem sample, 228 parts of uninfected leaf sample and 291 parts of uninfected phloem sample are collected together, and 1110 parts of sample are counted.
TABLE 1 sample number List
And 3, step 3: spectrum collection
The parameter resolution set by the spectrum scanning is 8cm-1The scanning speed was 5 frames/sec. The Labsolutions IR software is adopted for spectrum collection, a leaf or bark sample covers the lens of the detector, and the sample is covered by a light blocking cover of the instrument to avoid the influence of an external light source. In order to avoid the influence of the sampling position on the experimental result, the distribution of the spectral scanning points is shown in fig. 2, wherein fig. 2-1 and fig. 2-2 are the spectral scanning points of the blade: three points are taken along the main pulse, two ends (A, C) and a middle point (B), and a point (D, E) is taken at the left side and the right side of the leaf surface respectively, so that the straight line where the BDE three points are positioned is vertical to the main pulse; FIGS. 2-3 are spectral scan points of bark: five points are evenly taken along the phloem of the bark, so that the bark is evenly divided into 4 equal parts.
Main instrument and device
The spectrum acquisition device consists of a desktop computer and an IRTracer 100 NIR spectrometer, the spectrum analyzer is a Fourier near infrared spectrum system, the detector adopts an InGaAs detector, and the spectrum range is 12500-350 cm-1Spectral resolution of 0.25cm-1Spectral accuracy of 0.05cm-1Spectral accuracy 0.1cm-1。
And 4, step 4: spectral processing and data analysis
1.4.1 original Spectroscopy
For convenience of comparison, the absorbance of each raw spectrum was averagedThen subtracting to obtain the following differences D-Leaf, D-Bark, D-Hlb, D-Nhlb and D:
D=D-Leaf+D-Bark。
the original spectra are shown in fig. 3 and 4: the trends of the peak shape and peak position of the spectra collected from the leaf (FIG. 4) and bark phloem (FIG. 3) were similar, with the peak appearing at 6875cm-1Left and right and 5150cm-1And the left and right show that the obtained fruit tree sample spectrum has certain characteristics. Moreover, the spectral data of the infected sample and the non-infected sample have no obvious difference and cannot be distinguished by naked eyes.
Subtracting the average spectrum values of the unaffected and affected leaves to obtain a spectrum difference d 2; subtracting the average spectrum values of the uninfected barks and the infected leaves to obtain a spectrum difference value d 1; and D is set as the sum of the spectral differences of the bark and the leaves. The values of D1, D2 and D are magnified by 100 times and then plotted as a spectral difference chart, as shown in FIG. 5, where it can be directly seen that D < D-Bark < D-Leaf ≈ 0, i.e., the absorbance of the affected leaves is substantially equal to that of the uninfected leaves, while the absorbance of the affected Bark is greater than that of the uninfected Bark; and secondly, the D-Hlb is less than the D-Nhlb, and the difference between the absorbance of the infected leaf and the absorbance of the infected bark is reduced compared with that of a non-infected sample.
1.4.2 correction and prediction sets
Randomly dividing all samples into a correction set and a prediction set, wherein the specific classification is shown in table 2, the total correction set comprises 740 samples, wherein the leaf sample correction set comprises 395 samples, and the bark phloem sample correction set comprises 345 samples; the total prediction set totals 370 samples, wherein the leaf sample prediction set is 198, and the bark phloem sample prediction set is 172.
TABLE 2 number of samples in correction and prediction sets
1.4.3 data preprocessing
Performing data processing by using Unscrambler 10 software, respectively performing pretreatment on original data by using a Normalization method (Normalization), a standard normal distribution method (SNV), a multivariate scattering correction Method (MSC), a 1-order derivative method (First derivative) and a 2-order derivative method (Second derivative), modeling by using a Principal Component Regression (PCR) method, and comparing respective model effects, wherein the results are shown in table 3:
TABLE 3 prediction results of PCR model build after different pretreatments
It can be seen that the model after Normalization (Normalization) has the best predictive ability, and the correction set determines the coefficient R20.9390, the correction set root mean square error RMSEC is 2.0397 x 10-5Coefficient r is determined for prediction set20.9414, the prediction set root mean square error RMSEP is 1.8890 x 10-5。
Example 2
Randomly dividing all samples into a correction set and a prediction set as shown in Table 4, wherein the total correction set comprises 740 samples, the correction set of leaf samples comprises 395 samples, and the correction set of bark phloem samples comprises 345 samples; the total prediction set totals 370 samples, wherein the leaf sample prediction set is 198, and the bark phloem sample prediction set is 172.
TABLE 4 number of samples in correction and prediction sets
Preprocessing a leaf correction set, a bark phloem correction set and a total correction set by adopting a normalization method, and then respectively establishing models by adopting a PLSR method and a PCR method, wherein the establishment of the models respectively comprises the following steps: the near-infrared rapid detection method comprises a near-infrared rapid detection model for the navel orange yellow dragon disease based on a bark phloem, a near-infrared rapid detection model for the navel orange yellow dragon disease based on a leaf, and a near-infrared rapid detection model for the navel orange yellow dragon disease based on the bark phloem and the leaf.
Then, preprocessing the leaf prediction set, the bark phloem prediction set and the total prediction set by using a normalization method, and predicting each model respectively, wherein the prediction results are shown in table 5:
TABLE 5 prediction results of spectral modeling for different parts of navel orange
Both modeling methods show that: the precision of the three models is high (RMSEP is 10)-5Magnitude) and decreases in turn, while the decision coefficient increases in turn. Blade-based near-infrared rapid detection model for navel orange yellow dragon disease (RMSEPL. 1.6909X 0)-5,rL20.9396) the prediction accuracy is highest and the data fit is lowest. And a near-infrared rapid detection model (RMSEPC: 2.5676 × 0) for navel orange yellow dragon disease based on bark phloem and leaves-5,rC20.9603) with the lowest accuracy but the highest degree of data fit. The precision of the navel orange yellow dragon disease near infrared rapid detection model based on the bark phloem is only (RMSEPB: 1.8890 × 10)-5) Prediction ability (rB)20.9415) can be kept at a good level.
The invention provides a method for establishing a near-infrared detection model of navel orange yellow dragon disease, which is characterized in that a bark phloem is used as a sample, a Partial Least Squares Regression (PLSR) method is adopted to establish a near-infrared prediction model of navel orange yellow dragon disease based on the bark phloem after a data preprocessing method is optimized, and prediction accuracy and data fitting degree are considered.
The foregoing is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and decorations can be made without departing from the principle of the present invention, and these modifications and decorations should also be regarded as the protection scope of the present invention.
Claims (2)
1. A near-infrared detection model building method for navel orange yellow dragon disease based on bark phloem is characterized by comprising the following steps:
1) respectively taking navel orange trees with diseases and navel orange trees without diseases as standard fruit trees, and respectively stripping barks on lateral branches of the fruit trees as samples, wherein the width of each sample is 0.4-0.6 cm;
2) respectively taking 5 points on the phloem surface of the sample to perform spectrum scanning, and collecting spectra; the resolution of the scan is 8cm-1The scanning speed is 5 spectra/second; the spectrum range of the collected spectrum is 12500-350 cm-1Spectral resolution of 0.25cm-1Spectral accuracy of 0.05cm-1Spectral accuracy of 0.1cm-1;
3) Randomly dividing the collected spectral data into a correction set and a prediction set, preprocessing the data by using a normalization method, and establishing a partial least square regression model; before the data preprocessing, the method also comprises analyzing the acquired original spectrum, wherein the analysis comprises the step of calculating the average value of the absorbance of the spectrum by using the acquired original spectrumThen, calculating D-Bark, and obtaining a spectrogram by using the D-Bark;
2. the method as claimed in claim 1, wherein the step 2) of scanning the spectrum covers the sample over a lens of a detector and covers a light blocking cover over the sample.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811611126.7A CN109632699B (en) | 2018-12-27 | 2018-12-27 | Bark phloem-based near-infrared detection model establishment method for navel orange yellow dragon disease |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811611126.7A CN109632699B (en) | 2018-12-27 | 2018-12-27 | Bark phloem-based near-infrared detection model establishment method for navel orange yellow dragon disease |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109632699A CN109632699A (en) | 2019-04-16 |
CN109632699B true CN109632699B (en) | 2021-03-26 |
Family
ID=66078331
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811611126.7A Active CN109632699B (en) | 2018-12-27 | 2018-12-27 | Bark phloem-based near-infrared detection model establishment method for navel orange yellow dragon disease |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109632699B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113109327B (en) * | 2021-03-09 | 2023-11-17 | 杭州市林业科学研究院 | Method for predicting dry rot of hickory |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2013181558A1 (en) * | 2012-06-01 | 2013-12-05 | Agerpoint, Inc. | Systems and methods for monitoring agricultural products |
WO2015193885A2 (en) * | 2014-06-15 | 2015-12-23 | The State Of Israel, Ministry Of Agriculture & Rural Development, Agricultural Research Organization (Aro) (Volcani Center). | Method for huanglongbing (hlb) detection |
CN106018332A (en) * | 2016-07-21 | 2016-10-12 | 华南农业大学 | Near-infrared-spectrum citrus yellow shoot disease field detection method |
WO2018092135A1 (en) * | 2016-11-17 | 2018-05-24 | Fruitspec Ltd. | Method and system for crop yield estimation |
CN109115719A (en) * | 2018-10-08 | 2019-01-01 | 华东交通大学 | A kind of Citrus Huanglongbing pathogen Band fusion rapid detection method based on high light spectrum image-forming technology |
-
2018
- 2018-12-27 CN CN201811611126.7A patent/CN109632699B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2013181558A1 (en) * | 2012-06-01 | 2013-12-05 | Agerpoint, Inc. | Systems and methods for monitoring agricultural products |
WO2015193885A2 (en) * | 2014-06-15 | 2015-12-23 | The State Of Israel, Ministry Of Agriculture & Rural Development, Agricultural Research Organization (Aro) (Volcani Center). | Method for huanglongbing (hlb) detection |
CN106018332A (en) * | 2016-07-21 | 2016-10-12 | 华南农业大学 | Near-infrared-spectrum citrus yellow shoot disease field detection method |
WO2018092135A1 (en) * | 2016-11-17 | 2018-05-24 | Fruitspec Ltd. | Method and system for crop yield estimation |
CN109115719A (en) * | 2018-10-08 | 2019-01-01 | 华东交通大学 | A kind of Citrus Huanglongbing pathogen Band fusion rapid detection method based on high light spectrum image-forming technology |
Non-Patent Citations (5)
Title |
---|
Feasibility of Using the Optical Sensing Techniques for Early Detection of Huanglongbing in Citrus Seedlings;Alireza Pourreza等;《Robotics》;20170131;第6卷(第2期);全文 * |
PLS-DA Infrared Spectra Model of Citrus Leaves For The Characterization of Citrus Huanglongbing;Dong-mei CHEN等;《International Conference on Biomedical and Biological Engineering (BBE) 》;20160717;全文 * |
Use of near-infrared spectroscopyasan indicator of emerald ash borer infestation in white ash stem tissue;Kaelyn Finley等;《Forest Ecology and Management》;20160210;第366卷;摘要,第41页左栏,第44页左栏 * |
基于近红外技术柑橘黄龙病田间快速检测方法研究;饶敏等;《植物保护》;20171231;第43卷(第6期);全文 * |
柑桔黄龙病近红外光谱无损检测;刘燕德等;《刘燕德等》;20160731;第32卷(第14期);"0引言","1 材料与方法","2.2.2 偏最小二乘线性回归模型(PLS-DA)" * |
Also Published As
Publication number | Publication date |
---|---|
CN109632699A (en) | 2019-04-16 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Rodríguez-Pérez et al. | Evaluation of hyperspectral reflectance indexes to detect grapevine water status in vineyards | |
De Bei et al. | Non‐destructive measurement of grapevine water potential using near infrared spectroscopy | |
Liaghat et al. | Mid-infrared spectroscopy for early detection of basal stem rot disease in oil palm | |
Ourcival et al. | Exploring the relationships between reflectance and anatomical and biochemical properties in Quercus ilex leaves | |
Atkinson et al. | Discrimination between Betula pendula, Betula pubescens, and their hybrids using near-infrared reflectance spectroscopy | |
CN106092950A (en) | A kind of method differentiating dendrobium candidum medicinal material based near infrared spectrum | |
CN107328735A (en) | Rape species discrimination method based on terahertz light spectral technology | |
Lebot et al. | Application of near infrared reflectance spectroscopy for the evaluation of yam (Dioscorea alata) germplasm and breeding lines | |
Keskin et al. | Assessing nitrogen content of golf course turfgrass clippings using spectral reflectance | |
CN111562235A (en) | Method for rapidly identifying black-leaf outbreak disease and infection degree of tobacco leaves based on near infrared spectrum | |
CN109632699B (en) | Bark phloem-based near-infrared detection model establishment method for navel orange yellow dragon disease | |
Xie et al. | Spectral reflectance response to nitrogen fertilization in field grown corn | |
Yang | Nondestructive prediction of optimal harvest time of cherry tomatoes using VIS-NIR spectroscopy and PLSR calibration | |
Mat et al. | Prediction of sugarcane quality parameters using visible-shortwave near infrared spectroradiometer | |
CN109211802B (en) | Method for monitoring dead masson pine infected with pine wilt disease | |
CN102313715A (en) | Method for detecting honey quality base on laser technology | |
CN113176227A (en) | Method for rapidly predicting adulteration of dendrobium huoshanense in dendrobium hunan | |
CN113267458A (en) | Method for establishing quantitative prediction model of soluble protein content of sweet potatoes | |
Lin-Sheng et al. | Continuous wavelet analysis for diagnosing stress characteristics of leaf powdery mildew | |
CN109142237B (en) | Satellite spectral index for monitoring dead masson pine infected with pine wilt disease | |
CN109596577A (en) | The monitoring method that the construction method and wide angle of wheat powdery mildew state of illness monitoring model adapt to | |
CN104914053B (en) | Build trees canopy photosynthesis trait predictive model and detection Photosynthetic Characters | |
Krezhova et al. | Detection of biotic stress caused by apple stem grooving virus in apple trees using hyperspectral reflectance analysis | |
Bagheri et al. | Early detection of fire blight disease of pome fruit trees using visible-NIR spectrometry and dimensionality reduction methods | |
Ma et al. | Evaluation on intrinsic quality of Wolfiporia cocos influenced by cultivation methods by a multi-step IR macro-fingerprinting technique |
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 |