CN106841054B - Seed variety recognition methods and device - Google Patents
Seed variety recognition methods and device Download PDFInfo
- Publication number
- CN106841054B CN106841054B CN201710005012.7A CN201710005012A CN106841054B CN 106841054 B CN106841054 B CN 106841054B CN 201710005012 A CN201710005012 A CN 201710005012A CN 106841054 B CN106841054 B CN 106841054B
- Authority
- CN
- China
- Prior art keywords
- seed
- sample seed
- sample
- predetermined
- disaggregated model
- 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
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
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Theoretical Computer Science (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)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of seed variety recognition methods and devices, belong to image classification field.This method includes obtaining P high spectrum image of each sample seed under P wave band in test set;For each sample seed, characteristic parameter is obtained according to the P high spectrum image;The characteristic parameter is inputted into disaggregated model, obtains the prediction kind of each sample seed;Predetermined sample seed is selected from the test set according to the prediction kind, and the disaggregated model is updated according to the predetermined sample seed;The kind of sample seed described in the test set is identified using the updated disaggregated model;Solve the problems, such as when the seed of same breed due to time difference cause identify kind when accuracy it is not high;Reach and automatically updated disaggregated model, influence when reducing the time to identification seed variety improves the effect of the accuracy of identification kind.
Description
Technical field
The present embodiments relate to image classification field, in particular to a kind of seed variety recognition methods and device.
Background technique
Corn is one of staple crops of agricultural production, is the main source of grain, feed, fuel and the raw material of industry.
With the extensive use of hybrid seed technology, the kind of corn seed is more and more, is classified by the kind of corn seed
It becomes increasingly complex.
The advantages of combining machine vision and near infrared spectrum due to high light spectrum image-forming can reflect the inside of seed simultaneously
High light spectrum image-forming technology is widely used in seed classification in feature and surface.
However, the seed of the same breed in different year plantation, due to the tillage condition of different year, soil environment item
Part can change with weather conditions difference, the hyperspectral information of the seed of the same breed of different year, cause to utilize bloom
The accuracy for the seed variety that spectrogram picture and disaggregated model identify is not high.
Summary of the invention
In order to solve problems in the prior art, the embodiment of the invention provides a kind of seed variety method and devices.The skill
Art scheme is as follows:
In a first aspect, a kind of seed variety recognition methods is provided, this method comprises:
Obtain P high spectrum image of each sample seed under P wave band in test set;
For each sample seed, characteristic parameter is obtained according to the P high spectrum image;
The characteristic parameter is inputted into disaggregated model, obtains the prediction kind of each sample seed;
Predetermined sample seed is selected from the test set according to the prediction kind, and according to the predetermined sample kind
Son updates the disaggregated model;
The kind of sample seed described in the test set is identified using the updated disaggregated model.
Optionally, described that predetermined sample seed is selected from the test set according to the prediction kind, and according to institute
It states predetermined sample seed and updates the disaggregated model, comprising:
The sample seed is divided into several classifications according to the prediction kind;Sample seed in each classification
The prediction kind it is identical;
Calculate the class center of each classification;
Calculate the distance between each sample seed and the class center in each classification;
For each classification, the sample seed is arranged by the sequence of the distance from small to large;
Using the preceding n sample seed in each classification as the predetermined sample seed, and by the predetermined sample
The corresponding characteristic parameter of seed is deleted from whole characteristic parameters;
The predetermined sample seed is added into training set and obtains the new training set;
The disaggregated model is updated using the new training set;
Whether true detect predetermined condition;
If the predetermined condition is invalid, execute it is described by the characteristic parameter input disaggregated model, obtain each institute
The step of stating the prediction kind of sample seed;
Wherein, the predetermined condition isMore than or equal to predetermined threshold, alternatively, the predetermined condition is institute
It states number and reaches pre-determined number;Numj(i) it is determined when being i-th iteration and all the sample of j-th of classification when (i-1)-th iteration
The quantity of seed, Numj(i-1) be the i-th iteration when be confirmed as j-th of classification sample seed quantity.
Optionally, this method further include:
If the predetermined condition is set up, executes the updated disaggregated model of the utilization and identify the test set
Described in sample seed kind the step of.
Optionally, the distance between the sample seed calculated in each classification and the class center, comprising:
Utilize formulaCalculate the sample kind
The distance between sub and described class center;
Wherein,D is the dimension of vector.
It is optionally, described that characteristic parameter is obtained according to the P high spectrum image, comprising:
Obtain high spectrum image of the sample seed under predetermined band;
Image segmentation is carried out to the high spectrum image under predetermined band, the profile for obtaining the sample seed is bent
Line;
The contour curve is projected into the P wave band, P seed profile is obtained from the P high spectrum image
Image;
According to the seed contour images, the spectrum average feature of the corresponding sample seed of each wave band, institute are obtained
State the average value that spectrum average is characterized in the reflective light intensity of all pixels point in the seed contour images;
Using the corresponding P spectrum average feature of the P wave band as characteristic parameter.
Second aspect, provides a kind of seed variety identification device, which includes:
High spectrum image obtains module, for obtaining P EO-1 hyperion of each sample seed under P wave band in test set
Image;
Characteristic parameter obtains module, for being directed to each sample seed, is obtained according to the P high spectrum image special
Levy parameter;
It predicts that kind obtains module, for the characteristic parameter to be inputted disaggregated model, obtains each sample seed
Prediction kind;
Model modification module, for selecting predetermined sample seed from the test set according to the prediction kind, and
The disaggregated model is updated according to the predetermined sample seed;
Variety ecotype module, for identifying sample kind described in the test set using the updated disaggregated model
The kind of son.
Optionally, the model modification module, is specifically used for:
The sample seed is divided into several classifications according to the prediction kind;Sample seed in each classification
The prediction kind it is identical;
Calculate the class center of each classification;
Calculate the distance between each sample seed and the class center in each classification;
For each classification, the sample seed is arranged by the sequence of the distance from small to large;
Using the preceding n sample seed in each classification as the predetermined sample seed, and by the predetermined sample
The corresponding characteristic parameter of seed is deleted from whole characteristic parameters;
The predetermined sample seed is added into training set and obtains the new training set;
The disaggregated model is updated using the new training set;
Whether true detect predetermined condition;
If the predetermined condition is invalid, execute it is described by the characteristic parameter input disaggregated model, obtain each institute
The step of stating the prediction kind of sample seed;
Wherein, the predetermined condition isMore than or equal to predetermined threshold, alternatively, the predetermined condition is institute
It states number and reaches pre-determined number;Numj(i) it is determined when being i-th iteration and all the sample of j-th of classification when (i-1)-th iteration
The quantity of seed, Numj(i-1) be the i-th iteration when be confirmed as j-th of classification sample seed quantity.
Optionally, the model modification module, is also used to:
If the predetermined condition is set up, executes the updated disaggregated model of the utilization and identify the test set
Described in sample seed kind the step of.
Optionally, the model modification module, is also used to:
Utilize formulaCalculate the sample kind
The distance between sub and described class center;
Wherein,D is the dimension of vector.
Optionally, the characteristic parameter obtains module, is specifically used for:
Obtain high spectrum image of the sample seed under predetermined band;
Image segmentation is carried out to the high spectrum image under predetermined band, the profile for obtaining the sample seed is bent
Line;
The contour curve is projected into the P wave band, P seed profile is obtained from the P high spectrum image
Image;
According to the seed contour images, the spectrum average feature of the corresponding sample seed of each wave band, institute are obtained
State the average value that spectrum average is characterized in the reflective light intensity of all pixels point in the seed contour images;
Using the corresponding P spectrum average feature of the P wave band as characteristic parameter.
Technical solution provided in an embodiment of the present invention has the benefit that
By obtaining several sample seeds P high spectrum image under P wave band respectively in test set, according to bloom
Spectrogram picture obtains the characteristic parameter of each sample seed, obtains the prediction product of sample seed using characteristic parameter and disaggregated model
Kind is choosing predetermined sample seed using prediction kind, is updating disaggregated model according to predetermined sample seed, finally utilize after updating
Disaggregated model identification test set in whole seeds kind;The seed when same breed is solved since time difference causes to know
Accuracy not high problem when other kind;Reach and has automatically updated disaggregated model, shadow when reducing the time to identification seed variety
It rings, improves the effect of the accuracy of identification kind.
Detailed description of the invention
To describe the technical solutions in the embodiments of the present invention more clearly, make required in being described below to embodiment
Attached drawing is briefly described, it should be apparent that, drawings in the following description are only some embodiments of the invention, for
For those of ordinary skill in the art, without creative efforts, it can also be obtained according to these attached drawings other
Attached drawing.
Fig. 1 is a kind of flow chart of seed variety recognition methods shown according to an exemplary embodiment;
Fig. 2 is a kind of flow chart of the seed variety recognition methods shown according to another exemplary embodiment;
Fig. 3 A is a kind of partial schematic diagram of the high spectrum image shown according to another exemplary embodiment;
Fig. 3 B is a kind of partial schematic diagram of the high spectrum image shown according to another exemplary embodiment;
Fig. 3 C is a kind of partial schematic diagram of the contour curve of the sample seed shown according to another exemplary embodiment;
Fig. 3 D is a kind of seed profile diagram of sample seed for showing according to another exemplary embodiment under P wave band
Picture;
Fig. 3 E is a kind of spectrum average of the corresponding sample seed of each wave band of show according to another exemplary embodiment
The curve synoptic diagram of feature;
Fig. 4 is a kind of block diagram of the seed variety identification device shown according to another exemplary embodiment.
Specific embodiment
To make the object, technical solutions and advantages of the present invention clearer, below in conjunction with attached drawing to embodiment party of the present invention
Formula is described in further detail.
Referring to FIG. 1, it illustrates the flow charts of seed variety recognition methods provided by one embodiment of the present invention.Such as figure
Shown in 1, which be may comprise steps of:
Step 101, P high spectrum image of each sample seed under P wave band in test set is obtained.
Such as: test is concentrated with N number of sample seed, then gets N*P high spectrum image.
Specifically, whole sample seeds in test set are prevented setting in high spectrum image acquisition system, collects each
P high spectrum image of the sample seed under P wave band.
Optionally, the sample seed of test set China can be placed in high spectrum image acquisition system by several times, is acquired
P high spectrum image of each sample seed under P wave band.
It include the seed of several kinds in test set, the seed of each kind includes several times.
Step 102, for each sample seed, characteristic parameter is obtained according to P high spectrum image.
For each sample seed in test set, obtained according to P high spectrum image of the sample seed under P wave band
Take the sample seed for characteristic parameter.
One sample seed has a characteristic parameter, and the corresponding characteristic parameter of a sample seed is special by P spectrum average
Sign is constituted, and a sample seed has a corresponding spectrum average feature under a wave band.
Step 103, characteristic parameter is inputted into disaggregated model, obtains the prediction kind of each sample seed.
Preliminary classification model is generated according to training set, includes in the kind and test set of the seed for including in training set
Seed kind it is identical, in test set to the time of the seed of kind corresponding in training set difference.
Seed in training set includes multiple kinds, and the seed of each kind includes multiple times.Seed in test set
Kind when it is unknown.
Such as: seed in test set is A kind, B kind, C kind, the seed in training set be also A kind, B kind,
C kind, but the time of seed of A kind is different in the seed from training set of A kind in test set, B kind in test set
Seed is different from the time of the seed of B kind in training set, in test set in the seed and training set of C kind C kind seed
Time it is different.
Optionally, disaggregated model is least square method supporting vector machine disaggregated model.
Step 104, predetermined sample seed is selected from test set according to prediction kind, and more according to predetermined sample seed
New disaggregated model.
Predetermined sample seed is selected from whole sample seeds in test set according to prediction kind, what is selected is predetermined
The quantity of sample seed is less than the quantity of whole sample seeds in test set.
Optionally, predetermined sample seed can be repeatedly obtained from test set, the predetermined sample seed selected every time is not
Together;A predetermined sample seed is often selected, disaggregated model is once updated using the predetermined sample seed selected.
Step 105, the kind of sample seed in test set is identified using updated disaggregated model.
Optionally, the kind of whole sample seeds in test set is identified using updated disaggregated model, alternatively, utilizing
Updated disaggregated model identifies the kind of test concentrated part sample seed.
In conclusion seed variety recognition methods provided in an embodiment of the present invention, by obtaining each sample in test set
The seed P high spectrum image under P wave band respectively, the characteristic parameter of each sample seed is obtained according to high spectrum image,
The prediction kind of sample seed is obtained using characteristic parameter and disaggregated model, is choosing predetermined sample seed using prediction kind,
Disaggregated model is updated according to predetermined sample seed, finally utilizes the product of sample seed in updated disaggregated model identification test set
Kind;Solve the problems, such as when the seed of same breed due to time difference cause identify kind when accuracy it is not high;Reach certainly
Dynamic to update disaggregated model, when reducing the time to identification seed variety influence, improves the effect of the accuracy of identification kind.
Referring to FIG. 2, the flow chart of the seed variety recognition methods provided it illustrates another embodiment of the present invention.Such as
Shown in Fig. 2, which be may comprise steps of:
Step 201, P high spectrum image of each sample seed under P wave band in test set is obtained.
The step is expounded in a step 101, and which is not described herein again.
Step 202, for each sample seed, characteristic parameter is obtained according to P high spectrum image.
If there is N number of sample seed, N number of characteristic parameter is got.
Specifically, to each sample seed, characteristic parameter is obtained according to following steps 2021 to step 2025.
Step 2021, high spectrum image of the sample seed under predetermined band is obtained.
Optionally, predetermined band is the clearest wave band of high spectrum image.Optionally, predetermined band 782.59nm.
Step 2022, image segmentation is carried out to the high spectrum image under predetermined band, the profile for obtaining sample seed is bent
Line.
Optionally, before carrying out image segmentation to high spectrum image, figure enhancing is carried out to high spectrum image.
Optionally, image segmentation is carried out to the high spectrum image under predetermined band using threshold segmentation method, obtains sample
The contour curve of this seed.
Step 2023, contour curve is projected into P wave band, P seed profile diagram is obtained from P high spectrum image
Picture.
Using the method for outline projection, contour curve is projected under P wave band, from the corresponding high-spectrum of each wave band
A seed contour images are isolated as in.
The corresponding seed contour images of one wave band.One sample seed has P seed contour images.
Step 2024, according to seed contour images, the spectrum average feature of the corresponding sample seed of each wave band is obtained.
Spectrum average is characterized in the average value of the reflective light intensity of all pixels point in seed contour images.
According to the corresponding seed contour images of each wave band, the reflective light intensity of all pixels point in seed contour images is obtained
Average value.
One sample seed is corresponding with a spectrum average feature in a wave band, and a sample seed has P in P wave band
A spectrum average feature.
Step 2025, using the corresponding P spectrum average feature of P wave band as characteristic parameter.
Such as: P spectrum average feature of a sample seed is respectively p1、p2、p3、……、pp, then the sample seed
Characteristic parameter F=[p1,p2,p3,……,pp]。
Step 203, characteristic parameter is inputted into disaggregated model, obtains the prediction kind of each sample seed.
Optionally, characteristic parameter in sample seed is inputted into disaggregated model, obtains the prediction kind of each sample seed.
Step 204, sample seed is divided by several classifications according to prediction kind.
The prediction kind of sample seed in each classification is identical.
It should be noted that dividing class in this step if there is the unidentified sample seed for predicting kind out in step 203
When other, the unidentified sample seed for predicting kind out is not considered.
In each iteration, the quantity of the classification marked off can not be identical.
Step 205, the class center of each classification is calculated.
Assuming that the class center that T classification for T classification, is calculated in sample seed is [c1,...,ci,...,cT]。
Step 206, the distance between each sample seed and the class center in each classification are calculated.
The distance between each sample seed and class center in each classification are calculated using Pearson correlation coefficient.
Specifically, formula is utilizedCalculate sample
The distance between this seed and class center;
Wherein,Indicate a sample seed in each classification,Indicate the class center of each classification,D is the dimension of vector.
It should be noted thatValue it is bigger, illustrate indicate sample seed vector sum indicate class center
Vector it is more similar namely the distance between sample seed and class center are smaller.
Step 207, for each classification, sample seed is pressed into the sequence of distance from small to large and is arranged.
Step 208, using the preceding n sample seed in each classification as predetermined sample seed, and by predetermined sample seed
Corresponding characteristic parameter is deleted from whole characteristic parameters.
If the corresponding characteristic parameter of predetermined sample seed is deleted from whole characteristic parameters, in next iteration, suddenly
The predetermined sample seed of the predetermined sample seed namely deleted characteristic parameter that are slightly deleted characteristic parameter in test set is no longer joined
With the update of disaggregated model.
Step 209, predetermined sample seed is added into training set and obtains new training set.
Step 210, disaggregated model is updated using new training set.
Step 211, whether detection predetermined condition is true.
Wherein, predetermined condition isMore than or equal to predetermined threshold, alternatively, predetermined condition is that number reaches pre-
Determine number;Numj(i) be i-th iteration when and be all determined when (i-1)-th iteration j-th of classification sample seed quantity,
Numj(i-1) be i-th iteration when be confirmed as j-th of classification sample seed quantity.
Optionally, predetermined threshold is pre-set value.WhenWhen more than or equal to predetermined threshold, explanation
The prediction result of disaggregated model and the prediction result similitude of current iteration disaggregated model are very high after last iteration, namely classification
Model is not almost updated, needs to stop iteration.
Optionally, pre-determined number is maximum number of iterations.Maximum number of iterations and the sample seed obtained from test set
Quantity, the quantity for the seed variety for including in test set it is related.Optionally, if P high-spectrum in the case where obtaining P wave band
As when by several times obtain, then maximum number of iterations is also related with the quantity of sample seed obtained every time when obtaining image.
If predetermined condition is invalid, step 203 is held.
When predetermined condition is invalid, the disaggregated model in step 203 is the disaggregated model updated according to new training set,
Characteristic parameter after the corresponding characteristic parameter of deletion predetermined sample seed is inputted into disaggregated model.
If predetermined condition is set up, 212 are thened follow the steps.
Step 212, the kind of whole seeds in test set is identified using updated disaggregated model.
Optionally, the characteristic parameter for needing to identify the sample seed of kind in test set is obtained using high spectrum image, it will
It needs to identify that the characteristic parameter of the sample seed of kind inputs updated disaggregated model in test set, identifies sample seed
Kind.
Optionally, the sample seed for needing to identify kind is whole sample seeds in test set, alternatively, needing to identify product
The sample seed of kind is test concentrated part sample seed.
The sample seed for needing to identify can be the sample kind that characteristic parameter is deleted during updating disaggregated model
Son.
In conclusion seed variety recognition methods provided in an embodiment of the present invention, by obtaining each sample in test set
The seed P high spectrum image under P wave band respectively, the characteristic parameter of each sample seed is obtained according to high spectrum image,
The prediction kind of sample seed is obtained using characteristic parameter and disaggregated model, is choosing predetermined sample seed using prediction kind,
Disaggregated model is updated according to predetermined sample seed, finally utilizes the product of sample seed in updated disaggregated model identification test set
Kind;Solve the problems, such as when the seed of same breed due to time difference cause identify kind when accuracy it is not high;Reach certainly
Dynamic to update disaggregated model, when reducing the time to identification seed variety influence, improves the effect of the accuracy of identification kind.
In an illustrative example, it is assumed that the quantity of the sample seed of kind to be identified is 50 in test set, is obtained
P high spectrum image of each sample seed under P wave band in test set.50 seed specimens are obtained in 782.59nm wave band
Under high spectrum image, Fig. 3 A shows the partial schematic diagram of the high spectrum image;Image enhancement, figure are carried out to high spectrum image
3B shows the partial schematic diagram of enhanced high spectrum image;Using thresholding method, the profile of 50 sample seeds is obtained
Curve, Fig. 3 C show the partial schematic diagram of the contour curve of 50 sample seeds;The contour curve that will acquire is projected to P
Wave band obtains seed contour images of each sample seed under P wave band, as shown in Figure 3D;Further according to seed contour images,
The spectrum average feature for obtaining the corresponding sample seed of each wave band obtains each sample seed according to P spectrum average feature
Characteristic parameter, as shown in FIGURE 3 E.
Assuming that disaggregated model is least square method supporting vector machine disaggregated model Mlssvm, initial MlssvmIt is raw according to training set S
At.The characteristic parameter of 50 sample seeds is inputted into M respectivelylssvm, obtain the prediction label of each sample seed;Assuming that there is 47
Sample seed has identified prediction label, then is classified according to prediction label to 47 sample seeds, the sample of prediction label
Seed is assigned in the same classification, and prediction label has 4 classes, respectively A class, B class, C class, D class;Calculate separately out A class, B class, C
The class center of class, D class, then for each class calculate each sample to the class center of respective generic distance;It will be each
The nearest n sample seed in distance-like center can be expressed as predetermined sample seed, the set of predetermined sample seed in classification
Predetermined sample seed L is added into training set S, training set S=S+L by L;Predetermined sample seed in test set is ignored, accordingly
The characteristic parameter of predetermined sample seed is deleted on ground, and test set can be expressed as set U, U=U-L;It is updated using training set S
Mlssvm;It whether true detects predetermined condition, if predetermined condition is invalid, then executes the characteristic parameter of the sample seed in U
Input Mlssvm, the step of obtaining the prediction kind of each sample seed in U, until predetermined condition is set up, then using updated
MlssvmIdentify the kind of the sample seed in test set.
Following is apparatus of the present invention embodiment, can be used for executing embodiment of the present invention method.For apparatus of the present invention reality
Undisclosed details in example is applied, embodiment of the present invention method is please referred to.
Referring to figure 4., it illustrates the structural block diagrams of seed variety identification device provided by one embodiment of the present invention.
The seed variety identification device can be known by being implemented in combination with for software, hardware or both as the above-mentioned seed variety that can provide
The all or part of the terminal of other method.The device includes:
High spectrum image obtains module 410, P high under P wave band for obtaining each sample seed in test set
Spectrum picture;
Characteristic parameter obtains module 420, for being directed to each sample seed, obtains feature ginseng according to P high spectrum image
Number;
Prediction kind obtains module 430 and obtains the prediction of each sample seed for characteristic parameter to be inputted disaggregated model
Kind;
Model modification module 440, for selecting predetermined sample seed from test set according to prediction kind, and according to pre-
This seed of random sample updates disaggregated model;
Variety ecotype module 450, for identifying the kind of sample seed in test set using updated disaggregated model.
In conclusion seed variety identification device provided in an embodiment of the present invention, by obtaining each sample in test set
The seed P high spectrum image under P wave band respectively, the characteristic parameter of each sample seed is obtained according to high spectrum image,
The prediction kind of sample seed is obtained using characteristic parameter and disaggregated model, is choosing predetermined sample seed using prediction kind,
Disaggregated model is updated according to predetermined sample seed, finally utilizes the product of sample seed in updated disaggregated model identification test set
Kind;Solve the problems, such as when the seed of same breed due to time difference cause identify kind when accuracy it is not high;Reach certainly
Dynamic to update disaggregated model, when reducing the time to identification seed variety influence, improves the effect of the accuracy of identification kind.
Optionally, model modification module is specifically used for:
Sample seed is divided into several classifications according to prediction kind;The prediction kind phase of sample seed in each classification
Together;
Calculate the class center of each classification;
Calculate the distance between each sample seed and the class center in each classification;
For each classification, sample seed is pressed into the sequence of distance from small to large and is arranged;
Using the preceding n sample seed in each classification as predetermined sample seed, and by the corresponding spy of predetermined sample seed
Sign parameter is deleted from whole characteristic parameters;
Predetermined sample seed is added into training set and obtains new training set;
Disaggregated model is updated using new training set;
Whether true detect predetermined condition;
If predetermined condition is invalid, executes and characteristic parameter is inputted into disaggregated model, obtain the prediction of each sample seed
The step of kind;
Wherein, predetermined condition isMore than or equal to predetermined threshold, alternatively, predetermined condition is that number reaches pre-
Determine number;Numj(i) be i-th iteration when and be all determined when (i-1)-th iteration j-th of classification sample seed quantity,
Numj(i-1) be i-th iteration when be confirmed as j-th of classification sample seed quantity.
Optionally, model modification module is also used to:
If predetermined condition is set up, the kind that sample seed in test set is identified using updated disaggregated model is executed
The step of.
Optionally, model modification module is also used to:
Utilize formulaCalculate sample seed with
The distance between class center;
Wherein,D is the dimension of vector.
Optionally, characteristic parameter obtains module, is specifically used for:
Obtain high spectrum image of the sample seed under predetermined band;
Image segmentation is carried out to the high spectrum image under predetermined band, obtains the contour curve of sample seed;
Contour curve is projected into P wave band, P seed contour images are obtained from P high spectrum image;
According to seed contour images, the spectrum average feature of the corresponding sample seed of each wave band is obtained, spectrum average is special
Sign is the average value of the reflective light intensity of all pixels point in seed contour images;
Using the corresponding P spectrum average feature of P wave band as characteristic parameter.
It should be understood that seed variety identification computing device provided by the above embodiment is executing seed variety identification side
When method, only the example of the division of the above functional modules, in practical application, it can according to need and by above-mentioned function
Distribution is completed by different functional modules, i.e., the internal structure of equipment is divided into different functional modules, to complete above retouch
The all or part of function of stating.In addition, seed variety identification device provided by the above embodiment and seed variety recognition methods
Embodiment belongs to same design, and specific implementation process is detailed in embodiment of the method, and which is not described herein again.
The serial number of the above embodiments of the invention is only for description, does not represent the advantages or disadvantages of the embodiments.
Those of ordinary skill in the art will appreciate that realizing that all or part of the steps of above-described embodiment can pass through hardware
It completes, relevant hardware can also be instructed to complete by program, the program can store in a kind of computer-readable
In storage medium, storage medium mentioned above can be read-only memory, disk or CD etc..
The foregoing is merely presently preferred embodiments of the present invention, is not intended to limit the invention, it is all in spirit of the invention and
Within principle, any modification, equivalent replacement, improvement and so on be should all be included in the protection scope of the present invention.
Claims (5)
1. a kind of seed variety recognition methods, which is characterized in that the described method includes:
Obtain P high spectrum image of each sample seed under P wave band in test set;
For each sample seed, characteristic parameter is obtained according to the P high spectrum image;
The characteristic parameter is inputted into disaggregated model, obtains the prediction kind of each sample seed;
Predetermined sample seed is selected from the test set according to the prediction kind, and more according to the predetermined sample seed
The new disaggregated model;
The kind of sample seed described in the test set is identified using the updated disaggregated model;
Wherein, described that predetermined sample seed is selected from the test set according to the prediction kind, and according to described predetermined
Sample seed updates the disaggregated model, comprising:
The sample seed is divided into several classifications according to the prediction kind;The institute of sample seed in each classification
It is identical to state prediction kind;
Calculate the class center of each classification;
Calculate the distance between each sample seed and the class center in each classification;
For each classification, the sample seed is arranged by the sequence of the distance from small to large;
Using the preceding n sample seed in each classification as the predetermined sample seed, and by the predetermined sample seed
The corresponding characteristic parameter is deleted from whole characteristic parameters;
The predetermined sample seed is added into training set and obtains the new training set;
The disaggregated model is updated using the new training set;
Whether true detect predetermined condition;
If the predetermined condition is invalid, re-execute it is described by the characteristic parameter input disaggregated model, obtain each institute
The step of stating the prediction kind of sample seed and described predetermined sample is selected from the test set according to the prediction kind
Seed, and the step of disaggregated model is updated according to the predetermined sample seed;
The predetermined condition isMore than or equal to predetermined threshold, alternatively, the predetermined condition is described by institute for execution
State characteristic parameter input disaggregated model, the step of obtaining the prediction kind of each sample seed and described according to the prediction
Kind selects predetermined sample seed from the test set, and updates the disaggregated model according to the predetermined sample seed
The number of step reaches pre-determined number;Numj(i) j-th of classification is determined when repeating for i-th and all when (i-1)-th repetition
The quantity of sample seed, Numj(i-1) it is confirmed as the number of the sample seed of j-th of classification when repeating for the i-th
Amount.
2. the method according to claim 1, wherein the method also includes:
If the predetermined condition is set up, executes the updated disaggregated model of the utilization and identify institute in the test set
The step of stating the kind of sample seed.
3. the method according to claim 1, wherein the sample seed calculated in each classification and institute
State the distance between class center, comprising:
Utilize formulaCalculate the sample seed with
The distance between described class center;
Wherein,D is the dimension of vector.
4. method according to any one of claims 1 to 3, which is characterized in that described to be obtained according to the P high spectrum image
Characteristic parameter, comprising:
Obtain high spectrum image of the sample seed under predetermined band;
Image segmentation is carried out to the high spectrum image under predetermined band, obtains the contour curve of the sample seed;
The contour curve is projected into the P wave band, P seed contour images are obtained from the P high spectrum image;
According to the seed contour images, the spectrum average feature of the corresponding sample seed of each wave band, the light are obtained
Spectrum characteristics of mean is the average value of the reflective light intensity of all pixels point in the seed contour images;
Using the corresponding P spectrum average feature of the P wave band as characteristic parameter.
5. a kind of seed variety identification device, which is characterized in that described device requires 1 to 4 any described for perform claim
Method.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710005012.7A CN106841054B (en) | 2017-01-04 | 2017-01-04 | Seed variety recognition methods and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710005012.7A CN106841054B (en) | 2017-01-04 | 2017-01-04 | Seed variety recognition methods and device |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106841054A CN106841054A (en) | 2017-06-13 |
CN106841054B true CN106841054B (en) | 2019-06-07 |
Family
ID=59118401
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710005012.7A Active CN106841054B (en) | 2017-01-04 | 2017-01-04 | Seed variety recognition methods and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106841054B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111351766A (en) | 2020-02-27 | 2020-06-30 | 浙江大学 | Method for rapidly identifying identity of pumpkin seeds |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102072883A (en) * | 2010-07-07 | 2011-05-25 | 北京农业智能装备技术研究中心 | Device and method for detecting comprehensive quality of crop seeds |
CN102621077A (en) * | 2012-03-30 | 2012-08-01 | 江南大学 | Hyper-spectral reflection image collecting system and corn seed purity nondestructive detection method based on same |
CN105117734A (en) * | 2015-07-28 | 2015-12-02 | 江南大学 | Corn seed hyper-spectral image classification identification method based on model on-line updating |
CN105224960A (en) * | 2015-11-04 | 2016-01-06 | 江南大学 | Based on the corn seed classification hyperspectral imagery model of cognition update method of clustering algorithm |
CN106203452A (en) * | 2016-07-18 | 2016-12-07 | 江南大学 | Corn seed high spectrum image multiple features conversion method based on polyteny discriminant analysis |
CN106203522A (en) * | 2016-07-15 | 2016-12-07 | 西安电子科技大学 | Hyperspectral image classification method based on three-dimensional non-local mean filtering |
-
2017
- 2017-01-04 CN CN201710005012.7A patent/CN106841054B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102072883A (en) * | 2010-07-07 | 2011-05-25 | 北京农业智能装备技术研究中心 | Device and method for detecting comprehensive quality of crop seeds |
CN102621077A (en) * | 2012-03-30 | 2012-08-01 | 江南大学 | Hyper-spectral reflection image collecting system and corn seed purity nondestructive detection method based on same |
CN105117734A (en) * | 2015-07-28 | 2015-12-02 | 江南大学 | Corn seed hyper-spectral image classification identification method based on model on-line updating |
CN105224960A (en) * | 2015-11-04 | 2016-01-06 | 江南大学 | Based on the corn seed classification hyperspectral imagery model of cognition update method of clustering algorithm |
CN106203522A (en) * | 2016-07-15 | 2016-12-07 | 西安电子科技大学 | Hyperspectral image classification method based on three-dimensional non-local mean filtering |
CN106203452A (en) * | 2016-07-18 | 2016-12-07 | 江南大学 | Corn seed high spectrum image multiple features conversion method based on polyteny discriminant analysis |
Also Published As
Publication number | Publication date |
---|---|
CN106841054A (en) | 2017-06-13 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109584248B (en) | Infrared target instance segmentation method based on feature fusion and dense connection network | |
JP7058669B2 (en) | Vehicle appearance feature identification and vehicle search methods, devices, storage media, electronic devices | |
CN111027493B (en) | Pedestrian detection method based on deep learning multi-network soft fusion | |
US8379994B2 (en) | Digital image analysis utilizing multiple human labels | |
US20180373999A1 (en) | Targeted data augmentation using neural style transfer | |
CN108229531B (en) | Object feature extraction method and device, storage medium and electronic equipment | |
EP3532997B1 (en) | Training and/or using neural network models to generate intermediary output of a spectral image | |
CN111783779B (en) | Image processing method, apparatus and computer readable storage medium | |
CN110046598B (en) | Plug-and-play multi-scale space and channel attention remote sensing image target detection method | |
CN111967464B (en) | Weak supervision target positioning method based on deep learning | |
CN109472193A (en) | Method for detecting human face and device | |
CN109376757B (en) | Multi-label classification method and system | |
WO2022082848A1 (en) | Hyperspectral image classification method and related device | |
CN110826457B (en) | Vehicle detection method and device under complex scene | |
CN112529005B (en) | Target detection method based on semantic feature consistency supervision pyramid network | |
CN111401293A (en) | Gesture recognition method based on Head lightweight Mask scanning R-CNN | |
CN112001488A (en) | Training generative antagonistic networks | |
CN110363218A (en) | A kind of embryo's noninvasively estimating method and device | |
Choe et al. | The real-time mobile application for classifying of endangered parrot species using the CNN models based on transfer learning | |
CN106841054B (en) | Seed variety recognition methods and device | |
CN110210574B (en) | Synthetic aperture radar image interpretation method, target identification device and equipment | |
CN112132753A (en) | Infrared image super-resolution method and system for multi-scale structure guide image | |
CN116844032A (en) | Target detection and identification method, device, equipment and medium in marine environment | |
CN111178200A (en) | Identification method of instrument panel indicator lamp and computing equipment | |
CN109657083A (en) | The method for building up and device in textile picture feature library |
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 |