CN110082298A - A kind of wheat breed head blight comprehensive resistance identification method based on high spectrum image - Google Patents
A kind of wheat breed head blight comprehensive resistance identification method based on high spectrum image Download PDFInfo
- Publication number
- CN110082298A CN110082298A CN201910401015.1A CN201910401015A CN110082298A CN 110082298 A CN110082298 A CN 110082298A CN 201910401015 A CN201910401015 A CN 201910401015A CN 110082298 A CN110082298 A CN 110082298A
- Authority
- CN
- China
- Prior art keywords
- wheat
- resistance
- scab
- onset
- high 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.)
- Granted
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
- G01N21/17—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
- G01N21/27—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands using photo-electric detection ; circuits for computing concentration
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/188—Vegetation
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N2201/00—Features of devices classified in G01N21/00
- G01N2201/12—Circuits of general importance; Signal processing
- G01N2201/129—Using chemometrical methods
- G01N2201/1296—Using chemometrical methods using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/194—Terrestrial scenes using hyperspectral data, i.e. more or other wavelengths than RGB
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Immunology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Analytical Chemistry (AREA)
- Chemical & Material Sciences (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Pathology (AREA)
- Mathematical Physics (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Biochemistry (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Measuring Or Testing Involving Enzymes Or Micro-Organisms (AREA)
- Breeding Of Plants And Reproduction By Means Of Culturing (AREA)
Abstract
The invention discloses a kind of wheat breed head blight comprehensive resistance identification method based on high spectrum image, this method is using the different cultivars infection head blight germ wheat wheat head as research object, the knowledge of the numerous areas such as integrated use high light spectrum image-forming technology, spectroscopy, time series analysis, deep learning, proposes: wheat breed Resistance Identification method and wheat breed head blight comprehensive resistance identification method after the onset of for infecting non-period of disease wheat breed Resistance Identification method, being directed to.The present invention break through in difficult, the artificial identification operation of infection phase Resistance detecting that speed is slow, precision is low and chemical method in program it is complicated, at high cost the defects of, it is not only that the identification of kind wheat scab resistance provides the method for more fast accurate, and gibberellic hypha can be infected in wheat but non-period of disease carries out Resistance Identification to wheat breed.
Description
Technical field
The present invention relates to wheat scab detection and Resistance Identification, high light spectrum image-forming technology, time series analysis and depth
Learning algorithm field, specifically a kind of wheat breed head blight comprehensive resistance identification method based on high spectrum image.
Background technique
For wheat as one of three generalized grain of the world, wheat diseases problem becomes emphasis concerned by people.Wherein wheat is red
Mildew bacterium is to threaten a kind of maximum high plicae polonica to Wheat Production, this disease also known as red wheat head, reddish tone miasma, rotten wheat head, the wheat head
Withered, the wheat seed for infecting gibberellic hypha can generate mycotoxin based on deoxynivalenol in infection processs
(Xu Fei;Difference infects influence of the period to DON accumulation in wheat scab generation and seed), this will cause yield and seriously damages
It loses, and human health can also be seriously endangered by the mycotoxin that gibberellic hypha secretion generates, cause food-safety problem.Due to
Wheat yield caused by head blight is suffered heavy losses, especially head blight be very popular time wheat yield be up to 50% even can be exhausted
It produces.Moreover, wheat scab be currently in it is a kind of can anti-uncurable disease stage, once infection head blight, just centainly will cause not
The heavy losses that can be saved.It not can guarantee therapeutic effect not only even with pesticide at present, but also not environmentally, polluted significantly
Environment.Therefore, the problem of carrying out breeding to the wheat breed that disease-resistant type is presented in head blight, become the most important thing is selected.And it is small
The identification of wheat variety scab resistance is the important link of wheat breeding.
At this stage, the disease tassel yield and sick small ear that head blight is infected by wheat main for the identification of wheat scab resistance
Rate reflection, and statistics mainly is observed by plant protection personnel's artificial eye of profession for the detection of disease tassel yield and sick small ear rate,
This method takes time and effort incessantly, and judgement has subjectivity, can not precisely detect the disease tassel yield of wheat, and can not be to sense
The wheat for contaminating non-period of disease is diagnosed.Other than artificial detection, for the inspection of the DON toxin generated in wheat seed after the onset
It surveys, also occurs a variety of methods monitored with chemical-biological in succession now both at home and abroad, these chemical methodes study wheat flour, need
It will be first by wheat seed grinds, such as thin-layered chromatography (TLC), high performance liquid chromatography (HPLC) etc., although chemical method
Accurate quantitative analysis can be carried out to DON toxin, but these methods also have time-consuming, program complexity, inspection to a certain extent
Survey the problems such as at high cost.Other than complicated chemical method, there is researcher using hyperspectral technique to wheat at this stage
Head blight carries out grading diagnosis, and the side of different Threshold segmentations is extracted and used by the characteristic wave bands of principal component analysis twice
Method finally found that the grading effect of Local threshold segmentation method is best, and Grading accuracy rate reaches 90.78%.But local threshold point
Cutting method also has drawback, and this method is all the training mould by way of machine learning to the detection of the morbidity grade of wheat scab
Type, accuracy and simplicity are not so good as deep learning, and the noise problems such as the awn of wheat may be processed into lesion by Local threshold segmentation method
Position causes in different regions, and data have certain error, therefore can not be only completely suitable for typical classification algorithm
The higher data set of complexity.The network of depth can automatically extract out effective spectral signature in deep neural network, thus
Realize model construction end to end.It is most important yes, currently, for the research of wheat scab resistance being fallen ill by wheat
Rear disease tassel yield or classification situation compare, the detection infection germ but the case where not yet fall ill of having no idea.
Based on problem above, the present invention proposes a kind of wheat breed head blight comprehensive resistance identification based on high spectrum image
Method.Break through in difficult, the artificial identification operation of infection phase Resistance detecting that speed is slow, precision is low and chemical method in program it is multiple
The defects of miscellaneous, at high cost, only kind wheat scab resistance identification does not provide the method for more fast accurate, and can be
Wheat infects gibberellic hypha but non-period of disease carries out Resistance Identification to wheat breed.
Summary of the invention
In order to solve the problems in the prior art, the purpose of the present invention is to propose a kind of wheat breed based on high spectrum image
Head blight comprehensive resistance identification method;This method filled up wheat scab infect non-period of disease Resistance detecting blank and
Speed is slow in artificial Resistance Identification operation after the onset, precision is low and chemical method in program it is complicated, at high cost the defects of, cover
Wheat scab infect non-period of disease onset time detection and spectrum, characteristics of image variance analysis and after the onset grade from
Dynamic hierarchical detection, comprehensively considers the onset time for infect non-period of disease and illness after the onset is averaged severity, can it is comprehensive,
Fast and accurate realization wheat breed head blight comprehensive resistance identification.
To achieve the above object, the technical solution adopted by the present invention is as follows:
A kind of wheat breed head blight comprehensive resistance identification method based on high spectrum image, it is characterised in that including infection
Non- period of disease wheat breed Resistance Identification, after the onset wheat breed Resistance Identification and wheat breed based on above two result are red
The identification of mildew comprehensive resistance;Steps are as follows for specific identification method:
A, wheat samples acquire: choosing sample from wheat to be detected, and carry out respective markers according to kind;
B, the high spectrum image acquisition of the wheat wheat head: being imaged the wheat wheat head to be measured using Hyperspectral imager,
By obtained image, carries out the non-occurrent time resistance of wheat scab infection and compare and the detection of wheat scab grade after the onset
Model foundation;
C, wheat scab infects non-occurrent time resistance and compares: the step includes that different cultivars wheat scab just originates
The sick time detects the wheat scab resistance research of infection period identical with different cultivars wheat;Wherein, different cultivars wheat is red
The detection of mildew onset time is the rule and feature that spectrum changes with infection time after infecting germ by the exploration wheat head,
Timing key point is analyzed using dynamic time warping clustering algorithm based on sequential character, not based on timing EO-1 hyperion
It is detected with kind wheat scab onset time, the scab resistance of different cultivars wheat is compared with this;Different cultivars
The wheat scab resistance research of the identical infection period of wheat is the wheat head spectrum by verifying the identical infection period of different cultivars
Feature and characteristics of image variation compare the wheat of different cultivars based on spectral signature difference and characteristics of image variance analysis
Scab resistance;
D, wheat scab grade detection model is established after the onset: the wheat wheat head sample high-spectrum after the onset of will be complete
Picture is extracted, pretreatment, the selection of characteristic wave bands image by high spectrum image ROI, and the foundation of RCNN-VGG16 algorithm is combined to be based on
The wheat scab grade detection model of high spectrum image, and realize the target detection and mark of morbidity small ear and normal small ear,
Severity is averaged by different cultivars wheat scab illness to compare wheat breed scab resistance;
E, wheat breed head blight comprehensive resistance is identified: by onset time (in the resulting result of step c) and
Illness is averaged the comparison of severity (in the resulting result of step d) after the onset of different cultivars wheat, according to National agricultural industry
The average severity identification wheat resistance standard of standard NY/T1443.4-2007 wheat scab morbidity simultaneously combines onset time
Wheat Cultivars are specifically divided into the disease-resistant type of comprehensive resistance, anti-type in comprehensive resistance, sense type in comprehensive resistance, comprehensive resistance sense
Sick type.
Further, in stepb, the Hyperspectral imager includes CCD camera, Imspector spectrometer, mirror
Head, the linear halogen light source of 21V/150W, camera bellows and PC;High light spectrum image-forming wavelength band is in 358-1021nm.
Further, in step c and d, black and white correction is carried out to system, is just originated for different cultivars wheat scab
The detection of sick time is studied, the wheat wheat head high spectrum image after acquisition inoculation for 24 hours to morbidity cut-off completely after 21 days, and
It is marked by kind;For the analysis of the wheat head spectral signature and characteristics of image variation of the identical infection period of different cultivars
Research acquires different kinds 3 days after be inoculated with germ, 7 days, 14 days, 21 days high spectrum images, and mark infection time with
Kind;It is that all kind whole samples of acquisition are complete after 21 days for the foundation of wheat scab grade detection model after the onset
The wheat wheat head high spectrum image of morbidity.
Further, treatment research is carried out using the wheat wheat head high spectrum image every collected different cultivars for 24 hours,
By the high spectrum image from same sample, timing high spectrum image is constructed, corresponds to scab region when last morbidity completely, benefit
Region of interesting extraction is carried out with envi software, obtains timing spectrum, spectrum is pre-processed using convolution smoothing algorithm, and
To timing spectrum using successive projection algorithm carry out feature extraction, using dynamic time warping clustering algorithm timing spectroscopic data into
Row clustering, and sequential key point extraction is carried out, period of disease initial point is diagnosed;Based on having obtained for timing EO-1 hyperion
Different cultivars wheat scab onset time testing result, calculate the kind and be averaged onset time, compared with this
Compared with the scab resistance for infecting non-period of disease different cultivars wheat.
Further, using collected different kind 3 days after being inoculated with germ, 7 days, 14 days, 21 days EO-1 hyperions
Image carries out processing analysis, and the high spectrum image from same sample is carried out emerging to the sense of the whole wheat head using envi software
Interesting extracted region, by the average computation to region of interest pixel, the identical infection period of different cultivars wheat after being inoculated with
Average reflectance spectrum, observation spectrum figure compares SPECTRAL DIVERSITY on spectrogram.
Further, under observational characteristic wave band on high spectrum image, infected zone area, length, the difference of gray value,
Compare the non-period of disease different cultivars wheat scab resistance of infection by resulting SPECTRAL DIVERSITY and image difference.
Further, in step d, will fall ill the wheat wheat head sample high spectrum image after (inoculation 21 days after) completely,
Whole wheat head ROI region of interesting extraction is carried out to high spectrum image using envi software, SNV algorithm is recycled to carry out spectrum
Pretreatment, it is related to chlorophyll fluorescence parameters stabilization within the scope of visible light 520-680nm according to wheat chlorophyll reflectance spectrum,
Within this range, by the high spectrum image of naked-eye observation 650nm or so, the wheat picture of this wave band is chosen, and passes through mark
It is normal that software labelImg, which marks disease-free small ear, after ill small ear is sick, using based on five kinds of convolutional neural networks moulds
The Faster RCNN algorithm of type is trained, and establishes wheat ear scab infection etc. using RCNN-VGG16 algorithm based on this wave band
Grade identification model.
Further, the wheat picture input convolutional neural networks of selection obtain Feature Mapping, and Feature Mapping is inputted
Region suggests that network obtains the characteristic information of candidate frame and belongs to ill small ear or disease-free small ear by classifier differentiation, then
It is compared with the result of manual measurement, optimization algorithm;Again by the wheat ear scab grade detected, different cultivars is calculated
Illness is averaged severity, with determination after the onset of different cultivars scab resistance.
Further, in step e, in conjunction with the resulting onset time of step c and the resulting wheat breed hair of step d
The severity that is averaged of illness after being ill proposes wheat breed head blight comprehensive resistance identification method: i.e. when wheat breed morbidity is initial
It is the disease-resistant type product of comprehensive resistance that average severity of the time in 15-21 days and after the onset of, which is greater than 0 less than 2.0 wheat breed,
Kind;It is in 10-15 days and the sum of average severity after the onset of is less than greater than 2.0 when wheat breed initial time of falling ill
3.0 wheat breed is anti-type kind in comprehensive resistance;When wheat breed falls ill the initial time in 5-10 days and when morbidity
It is sense type kind in comprehensive resistance that average severity afterwards, which is greater than 3.0 wheat breed less than 3.5, when wheat breed morbidity is initial
Time be in 0-5 days and wheat breed of the average severity after the onset of greater than 3.5 is the susceptible type kind of comprehensive resistance.
Further, the average severity after the onset of wheat breed is calculated by the following formula:
Wherein, a: the kind primary sample number;B: the kind secondary sample number;C: the kind three-level number of samples;
D: the kind level Four number of samples;M: the total number of samples of the kind.
The present invention has the advantages that compared with prior art
(1) Breeding issue that wheat can be effectively solved to the research of wheat breed scab resistance, it is higher to select resistance
Breeding for quality, to reduce the hidden danger of wheat food safety.
(2) it is compared using the research to high spectrum image at present in artificial, chemical detection method, it can be lossless, quick, simple
Clean carries out scab resistance identification to wheat breed.
(3) research method of non-period of disease is infected to wheat scab using high spectrum image to resist for wheat breed head blight
Property research provide effective means, the blank of non-period of disease test variety resistance can not be infected in wheat scab by compensating for.
(4) it to the detection of wheat scab grade after the onset, is calculated in conjunction with EO-1 hyperion characteristic wave bands image using deep learning
Method can be handled a large amount of data set, realize model construction end to end, and the accurate precision of detection can be improved.
(5) it is combined the wheat scab that high spectrum image is obtained and infects non-period of disease morbidity initial time and morbidity
The illness severity that is averaged is put forward for the first time more comprehensive wheat breed head blight comprehensive resistance identification method afterwards.
To sum up, the present invention relates to high light spectrum image-forming technology, spectroscopy, time series analysis and deep learning algorithm and its
Field the relevant technologies take time and effort, program mainly for the height of testing cost present in the detection at this stage to wheat disease tassel yield
Complexity, and the problems such as the wheat wheat head sample fallen ill completely can only be detected.Wheat proposed by the present invention based on high spectrum image
Kind comprehensive resistance identification method, by being examined to the onset time for infecting non-period of disease different cultivars wheat high spectrum image
Survey the spectral signature of identical infection period and image characteristic analysis and utilization deep learning after the onset with different cultivars wheat
The wheat scab grade detection model based on high spectrum image of algorithm, when realizing that the morbidity of different cultivars wheat scab is initial
Between detection, the spectral signature of the identical infection period of different cultivars wheat, characteristics of image comparison in difference and grade after the onset
Automatic classification, and wheat scab morbidity initial time and the classification results of grade are calculated after the onset illness is combined to put down
Equal severity realizes the identification of wheat breed head blight comprehensive resistance.
Detailed description of the invention
Fig. 1 is total Technology Roadmap of the invention.
Fig. 2 is wheat morbidity initial time detection technique route map.
Fig. 3 is the identical infection period wheat scab resistance study route figure of different cultivars.
Fig. 4 is wheat scab automatic classification detection model Technology Roadmap.
Fig. 5 is each grade wheat head result figure of head blight;Wherein, Fig. 5 (a) is 0 grade of wheat head result figure, and Fig. 5 (b) is 1 grade of wheat
Fringe result figure, Fig. 5 (c) are 2 grades of wheat head result figures, and Fig. 5 (d) is 3 grades of wheat head result figures, and Fig. 5 (e) is 4 grades of wheat head result figures.
Specific embodiment
The present invention is described in detail with specific embodiment below in conjunction with the accompanying drawings.
(1) wheat samples acquisition and processing
Samples selection: the sample standard deviation in experiment comes from Nanjing Liuhe District experimental farm, and Wangshuibai, town wheat are taken in farm
168,22 4 kinds of Huaihe River wheat, Lip river wheat blooming stage health wheats, four kinds of wheat wheat heads are cultivated respectively, and carry out corresponding mark according to kind
Note.
Cultivate germ: 1. utilize potato dextrose agar (PDA) culture medium culture gibberellic hypha: potato cleans peeling
Stripping and slicing 400g, it is in soft shape that 5min is heated in boiling water to potato, and filtered juice is poured into pot and heated by filtering, and agar is added, stirs
It mixes addition glucose to continue to heat 5min, be sealed.2. germ isolates and purifies: the wheat seed of morbidity is obtained, by " rising
- 70% alcohol of mercury-aqua sterilisa " processing;0.3% grand celebration mycin is added in PDA culture medium, seed is inoculated in culture medium, people
Work constant temperature incubation purifies after 48h from colony edge picking mycelia, well develops into bacterium colony to mycelia, be transferred to the test tube slant PDA
On, adverse growth covers with PDA to mycelia after a couple of days, in preparation 1 × 105A/ml conidial suspension.
Inoculation experiments: the wheat wheat head sample that will be collected in the first step is all inoculated with, using syringe by spore suspension
It is injected into the single flower at the same position of the wheat head, and covers transparent plastic bag moisturizing.
(2) the high spectrum image acquisition of the wheat wheat head
Hyperspectral imager mainly includes CCD camera (Imperx company, the U.S.), Imspector spectrometer (Finland
Specim company), camera lens (German Schneider company), (U.S. Illumination is public for the linear halogen light source of 21V/150W
Department), camera bellows, the main components such as PC.High light spectrum image-forming wavelength band 358-1021nm.Clearly image information in order to obtain is more
Secondary adjusting pilot system parameter, factors and the single wheat such as research measure spectrum spectrum area and acquisition environment (light source, time, temperature etc.)
Relationship between fringe information, establishment can effectively carry sample characteristic information spectrum area and suitable spectral measurement mode and survey
Measure parameter.
Black and white correction is carried out to system, the detection for different cultivars wheat scab onset time is studied, every
Wheat wheat head high spectrum image to morbidity completely after 21 days after acquisition inoculation for 24 hours ends, and is marked by kind.For not
The analysis and research of the wheat head spectral signature and characteristics of image variation of infection period identical with kind, acquire different kinds and exist
It is inoculated with 3 days, 7 days, 14 days, 21 days high spectrum images after germ, and marks infection time and kind.It is red for wheat after the onset
The research of mildew grade detection model is established, and the wheat wheat head high-spectrum fallen ill completely after all kind whole samples 21 is acquired
Picture.
(3) head blight infects wheat disease tassel yield manual measurement
According to national standard GB/T 15796-2011 standard, the artificial rule that identifies combines agricultural production experience, to the wheat of acquisition
Wheat head sample is counted, according to infection small ear account for the ratio between total small ear of the wheat head, by sample according to infection level-one (ratio < 25%),
Second level (25%<ratio<50%), three-level (50%<ratio<75%), level Four (ratio>75%) carry out classification
(4) different times scab resistance detection model is infected based on high spectrum image
1. in order to establish different cultivars wheat scab onset time model, using every collected difference for 24 hours
The wheat wheat head high spectrum image of kind carries out treatment research.By the high spectrum image from same sample, timing bloom is constructed
Spectrogram picture corresponds to scab region when last morbidity completely, carries out area-of-interest (ROI) extraction using envi software, obtains
Timing spectrum pre-processes spectrum using convolution smoothing algorithm (SG), and utilizes successive projection algorithm to timing spectrum
(SPA) feature extraction is carried out, clustering is carried out using dynamic time warping (DTW) clustering algorithm timing spectroscopic data, goes forward side by side
Row sequential key point extracts, and diagnoses to period of disease initial point.Based at the beginning of the different cultivars wheat scab of timing EO-1 hyperion
Beginning disease time testing result is calculated the kind and is averaged onset time, the head blight of different cultivars wheat is compared with this
Resistance.Particular technique route map such as Fig. 2
2. in order to study the wheat head spectral signature of the identical infection period of different cultivars and characteristics of image variation, using adopting
The different kinds collected 3 days after being inoculated with germ, 7 days, 14 days, 21 days high spectrum images carry out treatment research.It will come from
The high spectrum image of same sample extract the area-of-interest (ROI) of the whole wheat head using envi software, by sense
The average computation of region of interest (ROI) pixel, the average reflectance spectrum of the identical infection period of different cultivars wheat after being inoculated with.
On observation spectrum figure, it is seen that in the wavelength of range different cultivars wheat identical infection period spectral reflectivity difference, due to
Head blight, the receipts of organic molecules principal structural component C-H, O-H, N-H key such as spectral information and protein, water, starch are infected
Contracting vibration is related to frequency multiplication, therefore SPECTRAL DIVERSITY can be compared on spectrogram.3 grades of frequencys multiplication of C-H be about 839nm, 918nm and
2 grades of frequencys multiplication of O-H and N-H are in 995nm or so.The Regional Representative's anthocyanidin and other pigments of about 600nm and 680nm.Observation
Difference, peak value in the wavelength band, valley, and analyze reason.Under observational characteristic wave band on high spectrum image, infected zone face
Product, length, the difference of gray value.Compare different cultivars wheat scab resistance by resulting SPECTRAL DIVERSITY and image difference.
Particular technique route map such as Fig. 3.
(5) the wheat scab grade detection model based on high spectrum image
Wheat wheat head sample high spectrum image after the onset of will be complete, carries out whole to high spectrum image using envi software
Wheat head ROI region of interesting extraction recycles SNV algorithm to pre-process spectrum, is existed according to wheat chlorophyll reflectance spectrum
It is related to chlorophyll fluorescence parameters stabilization within the scope of visible light 520-680nm, therefore within this range, it is arrived by naked-eye observation
The high spectrum image of 650nm or so is clearest, therefore has chosen the wheat picture of this wave band, and passes through marking software labelImg
Marking disease-free small ear is that normal uses the Faster based on five kinds of convolutional neural networks models after ill small ear is sick
RCNN algorithm is trained, and establishes wheat ear scab infection level identification model using RCNN-VGG16 algorithm based on this wave band.
Picture input convolutional neural networks obtain Feature Mapping, and Feature Mapping input area is suggested network (RPN) to obtain candidate frame
Characteristic information and ill small ear or disease-free small ear are belonged to by classifier differentiation, and carried out pair with the result of manual measurement
Than optimization algorithm.The test set for the infection wheat picture composition for containing 5 grades using 50 width is surveyed to the model come is trained
Examination, the model mAP value finally obtained are that 0.857, FPS (frame number of processing image per second) is 18, are classified total accuracy rate and reach
92%.Again by the wheat ear scab grade that detects, calculates different cultivars illness and be averaged severity, to determine different cultivars
Scab resistance.Particular technique route map such as Fig. 4.Each grade wheat head result figure such as Fig. 5 of head blight.
(6) wheat breed head blight comprehensive resistance identification method
The detection of the morbidity initial time of non-period of disease is infected in conjunction with above-mentioned wheat scab, and flat in illness after the onset
Equal severity proposes wheat breed head blight comprehensive resistance identification method, and this method specific requirement is when wheat breed morbidity is initial
Time be in 15-21 days and when after the onset of average severity be greater than 0 less than 2.0 wheat breed be the disease-resistant type of comprehensive resistance
Kind;It is in 10-15 days and the sum of average severity after the onset of is less than greater than 2.0 when wheat breed initial time of falling ill
3.0 wheat breed is anti-type kind in comprehensive resistance;When wheat breed falls ill the initial time in 5-10 days and when morbidity
It is sense type kind in comprehensive resistance that average severity afterwards, which is greater than 3.0 wheat breed less than 3.5, when wheat breed morbidity is initial
Time be in 0-5 days and wheat breed of the average severity after the onset of greater than 3.5 is the susceptible type kind of comprehensive resistance.Such as
The following table 1
15 < Time < 21&0 < average severity < 2.0 | The disease-resistant type kind of comprehensive resistance |
10 < Time < 15&2.0≤average severity < 3.0 | Anti- type kind in comprehensive resistance |
5 < Time < 10&3.0≤average severity < 3.5 | Sense type kind in comprehensive resistance |
0<Time<5& average severity>=3.5 | The susceptible type kind of comprehensive resistance |
1 wheat scab comprehensive resistance characterization and evaluation standard of table
(a: the kind primary sample number;B: the kind secondary sample number;
C: the kind three-level number of samples;D: the kind level Four number of samples;M: the total number of samples of the kind;)
The basic principles, main features and advantages of the invention have been shown and described above.Those skilled in the art
It should be appreciated that the protection scope that the above embodiments do not limit the invention in any form, all to be obtained using modes such as equivalent replacements
The technical solution obtained, falls in protection scope of the present invention.Part that the present invention does not relate to is the same as those in the prior art or can adopt
It is realized with the prior art.
Claims (10)
1. a kind of wheat breed head blight comprehensive resistance identification method based on high spectrum image, it is characterised in that including infecting not
Period of disease wheat breed Resistance Identification, after the onset wheat breed Resistance Identification and the wheat breed based on both the above result of study
The identification of head blight comprehensive resistance;Steps are as follows for specific identification method:
A, wheat samples acquire: choosing sample from wheat to be detected, and carry out respective markers according to kind;
B, the high spectrum image acquisition of the wheat wheat head: the wheat wheat head to be measured is imaged using Hyperspectral imager, is passed through
Obtained image carries out the non-occurrent time resistance of wheat scab infection and compares and wheat scab grade detection model after the onset
It establishes;
C, wheat scab infects non-occurrent time resistance and compares: when the step includes that different cultivars wheat scab is initially fallen ill
Between the wheat scab resistance of the identical with different cultivars wheat infection period of detection study;Wherein, different cultivars wheat scab
Onset time detection is the rule and feature that spectrum changes with infection time after infecting germ by the exploration wheat head, is based on
Sequential character analyzes timing key point using dynamic time warping clustering algorithm, the different product based on timing EO-1 hyperion
Kind wheat scab onset time detection, the scab resistance of different cultivars wheat is compared with this;Different cultivars wheat
The wheat scab resistance research of identical infection period is the wheat head spectral signature by verifying the identical infection period of different cultivars
With characteristics of image variation, the gibberella saubinetii of different cultivars is compared based on spectral signature difference and characteristics of image variance analysis
Sick resistance;
D, wheat scab grade detection model is established after the onset: the wheat wheat head sample high spectrum image after the onset of will be complete,
It is extracted by high spectrum image ROI, pretreatment, the selection of characteristic wave bands image, and RCNN-VGG16 algorithm is combined to establish based on height
The wheat scab grade detection model of spectrum picture, and realize the target detection and mark of morbidity small ear and normal small ear, lead to
It crosses different cultivars wheat scab illness and is averaged severity to compare wheat breed scab resistance;
E, wheat breed head blight comprehensive resistance is identified: passing through the illness after the onset of onset time and different cultivars wheat
The comparison of average severity, according to the average severity of National agricultural professional standard NY/T1443.4-2007 wheat scab morbidity
Identification wheat resistance standard simultaneously combines onset time that Wheat Cultivars are specifically divided into the disease-resistant type of comprehensive resistance, comprehensive anti-
Anti- type in property, sense type in comprehensive resistance, the susceptible type of comprehensive resistance.
2. a kind of wheat breed head blight comprehensive resistance identification method based on high spectrum image according to claim 1,
It is characterized in that, in stepb, the Hyperspectral imager include CCD camera, Imspector spectrometer, camera lens,
The linear halogen light source of 21V/150W, camera bellows and PC;High light spectrum image-forming wavelength band is in 358-1021nm.
3. a kind of wheat breed head blight comprehensive resistance identification method based on high spectrum image according to claim 1,
It is characterized in that, black and white correction is carried out to system, for different cultivars wheat scab onset time in step c and d
Detection research, wheat wheat head high spectrum image after acquisition inoculation for 24 hours to morbidity cut-off completely after 21 days, and press kind
It is marked;For the analysis and research of the wheat head spectral signature and characteristics of image variation of the identical infection period of different cultivars,
Different kinds is acquired 3 days after being inoculated with germ, 7 days, 14 days, 21 days high spectrum images, and marks infection time and kind;
For the foundation of wheat scab grade detection model after the onset, acquires all kind whole samples and fall ill completely after 21 days
Wheat wheat head high spectrum image.
4. a kind of wheat breed head blight comprehensive resistance identification side based on high spectrum image according to claim 1 or 3
Method, which is characterized in that the detection for different cultivars wheat scab onset time is studied, and is collected using every for 24 hours
The wheat wheat head high spectrum image of different cultivars carry out treatment research, by the high spectrum image from same sample, when building
Sequence high spectrum image, corresponds to scab region when last morbidity completely, carries out region of interesting extraction using envi software, obtains
Timing spectrum pre-processes spectrum using convolution smoothing algorithm, and carries out spy using successive projection algorithm to timing spectrum
Sign is extracted, and carries out clustering using dynamic time warping clustering algorithm timing spectroscopic data, and carry out sequential key point extraction,
Period of disease initial point is diagnosed;Obtained different cultivars wheat scab onset time based on timing EO-1 hyperion
Testing result calculates the kind and is averaged onset time, compares the red mould of the non-period of disease different cultivars wheat of infection with this
Sick resistance.
5. a kind of wheat breed head blight comprehensive resistance identification side based on high spectrum image according to claim 1 or 3
Method, which is characterized in that ground for the wheat head Spectral Characteristics Analysis and characteristics of image variance analysis of the identical infection period of different cultivars
Study carefully, using collected different kind 3 days after being inoculated with germ, 7 days, 14 days, 21 days high spectrum images carry out processing point
Analysis, the high spectrum image from same sample pass through the region of interesting extraction of the whole wheat head using envi software
To the average computation of region of interest pixel, the average reflectance spectrum of the identical infection period of different cultivars wheat after being inoculated with,
Observation spectrum figure, compares SPECTRAL DIVERSITY on spectrogram.
6. a kind of wheat breed head blight comprehensive resistance identification method based on high spectrum image according to claim 5,
It is characterized in that, under observational characteristic wave band on high spectrum image, infected zone area, length, the difference of gray value pass through gained
SPECTRAL DIVERSITY and image difference compare the non-period of disease different cultivars wheat scab resistance of infection.
7. a kind of wheat breed head blight comprehensive resistance identification method based on high spectrum image according to claim 1,
It is characterized in that, in step d, wheat scab grade detection model is established after the onset method particularly includes: will completely after the onset of
Wheat wheat head sample high spectrum image, whole wheat head ROI area-of-interest is carried out to high spectrum image using envi software and is mentioned
It takes, recycles SNV algorithm to pre-process spectrum, by the high spectrum image of naked-eye observation 650nm or so, choose this wave band
Wheat picture, and marking disease-free small ear by marking software labelImg is that normal uses base after ill small ear is sick
It is trained in the Faster RCNN algorithm of five kinds of convolutional neural networks models, RCNN-VGG16 algorithm is utilized based on this wave band
Establish wheat ear scab infection level identification model.
8. a kind of wheat breed head blight comprehensive resistance identification method based on high spectrum image according to claim 7,
It is characterized in that, the wheat picture input convolutional neural networks of selection obtain feature in Faster RCNN algorithm is trained
Feature Mapping input area suggestion network is obtained the characteristic information of candidate frame and is belonged to by classifier differentiation ill small by mapping
Fringe or disease-free small ear, then compare with the result of manual measurement, optimization algorithm;Wheat ear scab by detecting again
Grade calculates different cultivars illness and is averaged severity, with determination after the onset of different cultivars scab resistance.
9. a kind of wheat breed head blight comprehensive resistance identification method based on high spectrum image according to claim 1,
It is characterized in that, the identification of wheat breed head blight comprehensive resistance is initial time of being fallen ill by combining step c in step e
After the onset of testing result and step d illness severity as a result, i.e. when wheat breed fall ill the initial time be in 15-21 days and
It is the disease-resistant type kind of comprehensive resistance when average severity after the onset is greater than 0 less than 2.0 wheat breed;When wheat breed is fallen ill
The initial time is in 10-15 days and the sum of the average severity after the onset of is greater than 2.0 wheat breed less than 3.0 for synthesis
Anti- type kind in resistance;When wheat breed fall ill the initial time be in 5-10 days and ought after the onset of average severity be greater than
3.0 wheat breed less than 3.5 be comprehensive resistance in sense type kind, when wheat breed fall ill the initial time be in 0-5 days and
When wheat breed of the average severity after the onset greater than 3.5 is the susceptible type kind of comprehensive resistance.
10. according to claim 1 or a kind of wheat breed head blight comprehensive resistance identification side based on high spectrum image described in 9
Method, which is characterized in that the average severity after the onset of wheat breed is calculated by the following formula:
Wherein, a: the kind primary sample number;B: the kind secondary sample number;C: the kind three-level number of samples;D: should
Kind level Four number of samples;M: the total number of samples of the kind.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910401015.1A CN110082298B (en) | 2019-05-15 | 2019-05-15 | Hyperspectral image-based wheat variety gibberellic disease comprehensive resistance identification method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910401015.1A CN110082298B (en) | 2019-05-15 | 2019-05-15 | Hyperspectral image-based wheat variety gibberellic disease comprehensive resistance identification method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110082298A true CN110082298A (en) | 2019-08-02 |
CN110082298B CN110082298B (en) | 2020-05-19 |
Family
ID=67420191
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910401015.1A Active CN110082298B (en) | 2019-05-15 | 2019-05-15 | Hyperspectral image-based wheat variety gibberellic disease comprehensive resistance identification method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110082298B (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110596280A (en) * | 2019-10-29 | 2019-12-20 | 南京财经大学 | Rapid detection method for wheat vomitoxin pollution level based on hyperspectral image and spectrum information fusion |
CN110736750A (en) * | 2019-10-28 | 2020-01-31 | 安徽大学 | wheat scab detection method based on multi-angle field high-definition imaging |
CN110763698A (en) * | 2019-10-12 | 2020-02-07 | 仲恺农业工程学院 | Hyperspectral citrus leaf disease identification method based on characteristic wavelength |
CN112880734A (en) * | 2020-12-31 | 2021-06-01 | 中农新科(苏州)有机循环研究院有限公司 | Biological drying process digital monitoring system for reactor |
CN116523866A (en) * | 2023-04-26 | 2023-08-01 | 扬州大学 | Wheat scab resistance identification method, system, electronic equipment and storage medium |
CN116821698A (en) * | 2023-08-31 | 2023-09-29 | 中国科学技术大学 | Wheat scab spore detection method based on semi-supervised learning |
CN116973317A (en) * | 2023-06-13 | 2023-10-31 | 河北农业大学 | Method for identifying resistance of plant to plutella xylostella based on hyperspectral imaging technology |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102072885A (en) * | 2010-12-06 | 2011-05-25 | 浙江大学 | Machine vision-based paddy neck blast infection degree grading method |
CN103091284A (en) * | 2013-02-28 | 2013-05-08 | 河南工业大学 | Near-infrared spectrum technology based method for rapidly identifying wheat grains with FHB (Fusarium Head Blight) |
CN106777845A (en) * | 2017-03-22 | 2017-05-31 | 南京农业大学 | The method that sensitive parameter builds wheat leaf blade powdery mildew early monitoring model is extracted based on subwindow rearrangement method |
CN107527326A (en) * | 2017-08-17 | 2017-12-29 | 安徽农业大学 | A kind of wheat scab diagnostic method based on high light spectrum image-forming |
CN107576618A (en) * | 2017-07-20 | 2018-01-12 | 华南理工大学 | Pyricularia Oryzae detection method and system based on depth convolutional neural networks |
CN109035231A (en) * | 2018-07-20 | 2018-12-18 | 安徽农业大学 | A kind of detection method and its system of the wheat scab based on deep-cycle |
CN109187552A (en) * | 2018-08-30 | 2019-01-11 | 安徽农业大学 | A kind of gibberella saubinetii disease grade determination method based on cloud model |
CN109544538A (en) * | 2018-11-27 | 2019-03-29 | 安徽大学 | Wheat scab disease grade grading method and device |
CN109657653A (en) * | 2019-01-21 | 2019-04-19 | 安徽大学 | A kind of wheat seed head blight recognition methods based on Imaging Hyperspectral Data |
-
2019
- 2019-05-15 CN CN201910401015.1A patent/CN110082298B/en active Active
Patent Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102072885A (en) * | 2010-12-06 | 2011-05-25 | 浙江大学 | Machine vision-based paddy neck blast infection degree grading method |
CN103091284A (en) * | 2013-02-28 | 2013-05-08 | 河南工业大学 | Near-infrared spectrum technology based method for rapidly identifying wheat grains with FHB (Fusarium Head Blight) |
CN103091284B (en) * | 2013-02-28 | 2015-07-15 | 河南工业大学 | Near-infrared spectrum technology based method for rapidly identifying wheat grains with FHB (Fusarium Head Blight) |
CN106777845A (en) * | 2017-03-22 | 2017-05-31 | 南京农业大学 | The method that sensitive parameter builds wheat leaf blade powdery mildew early monitoring model is extracted based on subwindow rearrangement method |
CN107576618A (en) * | 2017-07-20 | 2018-01-12 | 华南理工大学 | Pyricularia Oryzae detection method and system based on depth convolutional neural networks |
CN107527326A (en) * | 2017-08-17 | 2017-12-29 | 安徽农业大学 | A kind of wheat scab diagnostic method based on high light spectrum image-forming |
CN109035231A (en) * | 2018-07-20 | 2018-12-18 | 安徽农业大学 | A kind of detection method and its system of the wheat scab based on deep-cycle |
CN109187552A (en) * | 2018-08-30 | 2019-01-11 | 安徽农业大学 | A kind of gibberella saubinetii disease grade determination method based on cloud model |
CN109544538A (en) * | 2018-11-27 | 2019-03-29 | 安徽大学 | Wheat scab disease grade grading method and device |
CN109657653A (en) * | 2019-01-21 | 2019-04-19 | 安徽大学 | A kind of wheat seed head blight recognition methods based on Imaging Hyperspectral Data |
Non-Patent Citations (6)
Title |
---|
JIANBO QIU ET AL: "Fusarium Toxins in Chinese Wheat since the 1980s", 《MDPI》 * |
STEPHEN R. DELWICHE ET AL: "Hyperspectral imaging for detection of scab in wheat", 《SPIE》 * |
中华人民共和国农业部: "小麦抗虫性评价技术规范第4部分:小麦抗赤霉病评价技术规范", 《中华人民共和国农业行业标准NY/T 1443.4-2007》 * |
梁琨等: "基于高光谱成像技术的小麦籽粒赤霉病识别", 《农业机械学报》 * |
金秀等: "基于动态时间弯曲的马铃薯干时序高光谱诊断方法腐病发病期", 《食品科学》 * |
黄明祥等: "松材线虫病害高光谱时序与敏感特征研究", 《遥感技术与应用》 * |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110763698A (en) * | 2019-10-12 | 2020-02-07 | 仲恺农业工程学院 | Hyperspectral citrus leaf disease identification method based on characteristic wavelength |
CN110736750A (en) * | 2019-10-28 | 2020-01-31 | 安徽大学 | wheat scab detection method based on multi-angle field high-definition imaging |
CN110736750B (en) * | 2019-10-28 | 2022-03-04 | 安徽大学 | Wheat scab detection method based on multi-angle field high-definition imaging |
CN110596280A (en) * | 2019-10-29 | 2019-12-20 | 南京财经大学 | Rapid detection method for wheat vomitoxin pollution level based on hyperspectral image and spectrum information fusion |
CN110596280B (en) * | 2019-10-29 | 2022-04-05 | 南京财经大学 | Rapid detection method for wheat vomitoxin pollution level based on hyperspectral image and spectrum information fusion |
CN112880734A (en) * | 2020-12-31 | 2021-06-01 | 中农新科(苏州)有机循环研究院有限公司 | Biological drying process digital monitoring system for reactor |
CN116523866A (en) * | 2023-04-26 | 2023-08-01 | 扬州大学 | Wheat scab resistance identification method, system, electronic equipment and storage medium |
CN116523866B (en) * | 2023-04-26 | 2023-12-01 | 扬州大学 | Wheat scab resistance identification method, system, electronic equipment and storage medium |
CN116973317A (en) * | 2023-06-13 | 2023-10-31 | 河北农业大学 | Method for identifying resistance of plant to plutella xylostella based on hyperspectral imaging technology |
CN116821698A (en) * | 2023-08-31 | 2023-09-29 | 中国科学技术大学 | Wheat scab spore detection method based on semi-supervised learning |
CN116821698B (en) * | 2023-08-31 | 2024-01-05 | 中国科学技术大学 | Wheat scab spore detection method based on semi-supervised learning |
Also Published As
Publication number | Publication date |
---|---|
CN110082298B (en) | 2020-05-19 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110082298A (en) | A kind of wheat breed head blight comprehensive resistance identification method based on high spectrum image | |
CN108444928B (en) | Method for identifying cereal seed frostbite condition by using seed embryo spectrogram characteristic wave band | |
CN102495005B (en) | Method for diagnosing crop water deficit through hyperspectral image technology | |
CN110132862B (en) | Construction method and application of disease index special for wheat scab detection | |
Huang et al. | Detection of insect-damaged vegetable soybeans using hyperspectral transmittance image | |
Dey et al. | An analysis of leaf chlorophyll measurement method using chlorophyll meter and image processing technique | |
CN110763698B (en) | Hyperspectral citrus leaf disease identification method based on characteristic wavelength | |
Rios et al. | Development and validation of a standard area diagram set to assess blast severity on wheat leaves | |
CN105067532B (en) | A kind of method for differentiating sclerotinia sclerotiorum and gray mold early stage scab | |
CN101881726A (en) | Nondestructive detection method for comprehensive character living bodies of plant seedlings | |
CN109948596A (en) | A method of rice identification and crop coverage measurement are carried out based on vegetation index model | |
CN104697943A (en) | Rape water stress nondestructive testing method | |
Bensaeed et al. | Oil palm fruit grading using a hyperspectral device and machine learning algorithm | |
KR102095539B1 (en) | Method for measuring growth amount by image analyzing ginseng | |
CN112287886B (en) | Wheat plant nitrogen content estimation method based on hyperspectral image fusion map features | |
CN109765194A (en) | Fructus lycii Production area recognition method based on high light spectrum image-forming technology | |
CN110363125A (en) | Using the method for Model Transfer identification different cultivars Citrus Huanglongbing pathogen | |
Pujari et al. | Recognition and classification of produce affected by identically looking powdery mildew disease | |
CN106332713A (en) | Method for identifying early-phase bruise of loquat through SD-OCT image | |
Lisu et al. | Rapid diagnosis of nitrogen nutrition status in rice based on static scanning and extraction of leaf and sheath characteristics | |
US20230095405A1 (en) | Method and system for screening spectral indexes of rice resistant to bacterial blight | |
CN109470639A (en) | Diaphania larva based on visible light and near-infrared high light spectrum image-forming and its method for quickly identifying to mulberry leaf damage | |
CN109389581A (en) | The method of wheat leaf rust classification based on infrared thermal imaging edge detection process | |
Prasher et al. | Analysis of DenseNet201 with SGD optimizer for diagnosis of multiple rice leaf diseases | |
CN114965346A (en) | Kiwi fruit quality detection method based on deep learning and hyperspectral imaging technology |
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 | ||
CB02 | Change of applicant information | ||
CB02 | Change of applicant information |
Address after: 210031 No. 40, Taiwan Road, Pukou District, Nanjing, Jiangsu Applicant after: NANJING AGRICULTURAL University Address before: 211225, Jiangsu, Nanjing province Lishui Baima Town National Agricultural Science and Technology Park, Nanjing Agricultural University base Applicant before: NANJING AGRICULTURAL University |
|
GR01 | Patent grant | ||
GR01 | Patent grant |