CN109344917A - A kind of species identification method and identification system of Moth insects - Google Patents
A kind of species identification method and identification system of Moth insects Download PDFInfo
- Publication number
- CN109344917A CN109344917A CN201811324618.8A CN201811324618A CN109344917A CN 109344917 A CN109344917 A CN 109344917A CN 201811324618 A CN201811324618 A CN 201811324618A CN 109344917 A CN109344917 A CN 109344917A
- Authority
- CN
- China
- Prior art keywords
- image
- module
- insect
- euproctis
- acquisition
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 43
- 241000894007 species Species 0.000 title claims abstract description 29
- 241000238631 Hexapoda Species 0.000 title claims abstract description 20
- 239000002574 poison Substances 0.000 claims abstract description 55
- 231100000614 poison Toxicity 0.000 claims abstract description 55
- 241001331999 Euproctis Species 0.000 claims abstract description 43
- 238000012850 discrimination method Methods 0.000 claims abstract description 17
- 238000001514 detection method Methods 0.000 claims abstract description 14
- 239000000284 extract Substances 0.000 claims abstract description 10
- 230000005055 memory storage Effects 0.000 claims abstract description 8
- 238000012545 processing Methods 0.000 claims description 36
- 238000004422 calculation algorithm Methods 0.000 claims description 24
- 230000006870 function Effects 0.000 claims description 19
- 238000012549 training Methods 0.000 claims description 12
- 238000013528 artificial neural network Methods 0.000 claims description 11
- 238000000605 extraction Methods 0.000 claims description 9
- 210000002569 neuron Anatomy 0.000 claims description 9
- 230000005540 biological transmission Effects 0.000 claims description 8
- 238000005260 corrosion Methods 0.000 claims description 6
- 230000000877 morphologic effect Effects 0.000 claims description 5
- 238000010276 construction Methods 0.000 claims description 4
- 108090000623 proteins and genes Proteins 0.000 claims description 4
- 238000012897 Levenberg–Marquardt algorithm Methods 0.000 claims description 3
- 230000007797 corrosion Effects 0.000 claims description 3
- 238000003709 image segmentation Methods 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 238000000513 principal component analysis Methods 0.000 claims description 3
- 239000012141 concentrate Substances 0.000 claims description 2
- 230000004927 fusion Effects 0.000 description 12
- 239000011159 matrix material Substances 0.000 description 9
- 238000003860 storage Methods 0.000 description 8
- 238000007781 pre-processing Methods 0.000 description 7
- 230000008569 process Effects 0.000 description 6
- 230000009466 transformation Effects 0.000 description 6
- 238000011156 evaluation Methods 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 238000002156 mixing Methods 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 238000004590 computer program Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 241001465754 Metazoa Species 0.000 description 2
- 238000010521 absorption reaction Methods 0.000 description 2
- 238000000354 decomposition reaction Methods 0.000 description 2
- 235000013399 edible fruits Nutrition 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 238000007500 overflow downdraw method Methods 0.000 description 2
- 229910052704 radon Inorganic materials 0.000 description 2
- SYUHGPGVQRZVTB-UHFFFAOYSA-N radon atom Chemical compound [Rn] SYUHGPGVQRZVTB-UHFFFAOYSA-N 0.000 description 2
- 239000007787 solid Substances 0.000 description 2
- 241000254032 Acrididae Species 0.000 description 1
- 241000256844 Apis mellifera Species 0.000 description 1
- 241000238421 Arthropoda Species 0.000 description 1
- 241000894006 Bacteria Species 0.000 description 1
- 241001674044 Blattodea Species 0.000 description 1
- 108010000912 Egg Proteins Proteins 0.000 description 1
- 102000002322 Egg Proteins Human genes 0.000 description 1
- 241000510032 Ellipsaria lineolata Species 0.000 description 1
- 241000233866 Fungi Species 0.000 description 1
- 241000721703 Lymantria dispar Species 0.000 description 1
- 241000238633 Odonata Species 0.000 description 1
- 241000700605 Viruses Species 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000012271 agricultural production Methods 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000009412 basement excavation Methods 0.000 description 1
- 230000000903 blocking effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000013499 data model Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000003628 erosive effect Effects 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 230000014759 maintenance of location Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 210000004681 ovum Anatomy 0.000 description 1
- 238000002203 pretreatment Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000004445 quantitative analysis Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 239000004576 sand Substances 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 230000007480 spreading Effects 0.000 description 1
- 238000003892 spreading Methods 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
- 238000009827 uniform distribution Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- 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
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/06—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons
- G06N3/061—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons using biological neurons, e.g. biological neurons connected to an integrated circuit
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/155—Segmentation; Edge detection involving morphological operators
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/187—Segmentation; Edge detection involving region growing; involving region merging; involving connected component labelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20036—Morphological image processing
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Biophysics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Evolutionary Biology (AREA)
- Molecular Biology (AREA)
- Neurology (AREA)
- Microelectronics & Electronic Packaging (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
Abstract
The invention belongs to insect species authentication technique fields, disclose the species discrimination method and identification system of a kind of Euproctis insect；It is entomologized the image data information of front and rear wing using image pick-up device；Detection identification is carried out using image of the image processing software to acquisition, extracts poison moth wing image math-morphological features, main control module analyzes math-morphological features, filters out the characteristic parameter for Classification and Identification；Judgement is compared using network data base downloading Euproctis insect poison moth wing image data information；And big data resource is concentrated to handle the image information of acquisition using Cloud Server by cloud service module；The poison moth wing image data information of memory storage acquisition is utilized by data memory module；And the poison moth wing image data information of display display acquisition is utilized by display module.The present invention can solve conventional sorting methods labor intensive, time and the bad problem of efficiency；Identity card is established for species, accurately distinguishes allied species.
Description
Technical field
The invention belongs to insect species authentication technique field more particularly to a kind of species discrimination methods of Euproctis insect.
Background technique
Caste is various, comes in every shape, and belongs to the arthropod in invertebrate, and being that quantity is most on the earth moves
Object group accounts in all biological species (including bacterium, fungi, virus) more than 50%, their trace almost spreads over generation
Each corner on boundary.At the beginning of 21 century, insect known to the mankind has more than 100 ten thousand kinds, but still waits to find there are many type.
Insect most species in the animal kingdom, quantity is maximum, causes significant impact to agricultural production and human health.It is most common to have locust
Worm, butterfly, honeybee, dragonfly, fly, grasshopper, cockroach etc..Not only type is more for insect, and individual amount of the same race is also very frightened
People.The distribution face of insect is wide, can be in contrast, almost spreading the entire earth without the animal of other guiding principles.Point have different
Type, most insects can do sample, be the utilizable good biological resource of the mankind.However, the object of existing Euproctis insect
Kind, which identifies, to be needed to be judged again by the insect under eye-observation microscope, therefore also all there are conventional sorting methods to expend
Manpower, time and the bad problem of efficiency.Meanwhile Euproctis insect is identified or dependence adult morphological method, on ship
What is be frequently found is larva and ovum, is difficult precise Identification with morphological method, this, which works to the inspection and quarantine of port gypsymoth, increases
Difficulty is added.In the prior art, the image processing algorithm that video camera uses reduces original image procossing precision, for mass center
Required precision is not achieved in method and Gauss curve fitting method, image zooming-out precision；In conventional images display technology, local noise is to aobvious
The clarity operation result of diagram picture has a great impact, and reduces the clarity that image is shown；Existing data search simultaneously
In the process, algorithm is complicated, and convergence rate is slow, expansible and low efficiency.
In conclusion problem of the existing technology is:
In the prior art, the image processing algorithm that video camera uses reduces original image procossing precision, for centroid method
With Gauss curve fitting method, required precision is not achieved in image zooming-out precision.
In conventional images display technology, local noise has a great impact to the clarity operation result of display image, drops
The clarity that low image is shown.
During existing data search, algorithm is complicated, and convergence rate is slow, expansible and low efficiency.
The process that the prior art identifies poison moth, low efficiency and there are inaccuracies.
The processing that the prior art cannot refine the poison moth wing image math-morphological features of acquisition causes to exist to a certain degree
Driscrimination error.
During the Euproctis insect poison moth wing image data information of prior art storage acquisition, there are many storages
Demand increases Algorithms T-cbmplexity.
Summary of the invention
In view of the problems of the existing technology, the present invention provides a kind of species discrimination methods of Euproctis insect.
The invention is realized in this way a kind of species discrimination method of Euproctis insect, the species of the Euproctis insect
Discrimination method includes:
The first step is entomologized the image data information of front and rear wing using image pick-up device；
Second step carries out detection identification using image of the image processing software to acquisition, extracts poison moth wing image mathematics shape
State feature, main control module analyze math-morphological features, filter out the characteristic parameter for Classification and Identification；
Third step compares judgement using network data base downloading Euproctis insect poison moth wing image data information；And
Big data resource is concentrated to handle the image information of acquisition using Cloud Server by cloud service module；
4th step utilizes the poison moth wing image data information of memory storage acquisition by data memory module；And pass through
The poison moth wing image data information that display module is acquired using display display.
Further, poison moth wing image math-morphological features are extracted, math-morphological features are analyzed, are filtered out for dividing
The characteristic parameter of class identification, using BP neural network classifier, the method for building up of BP neural network classifier, including following step
It is rapid:
Step 1, using 3 layers of BP neural network of a hidden layer, input layer number 5 is corresponding through principal component analysis
The F of foundation1、F2、F3、F4And F5Variable；
Step 2, output layer neuron number are that different moth category insects is carried out assignment by corresponding moth category caste；
Step 3, hidden layer neuron number are set to variable, are calculated according to following empirical equation；
In formula: InIndicate input layer number, OnIndicate output layer neuron number, a ∈ [1,10]；
Step 4, input layer to hidden layer use tansig transmission function；Hidden layer to output layer transmission function be purelin；
Training algorithm uses non-linear damped least-square method optimization algorithm, function trainlm；
Step 5, training error maximum cycle are 2000, network training result and training sample data relative error
Control is 1 × 10-6Within, it is believed that reach required precision, learning coefficient 0.05.
Further, poison moth wing image math-morphological features are extracted and use Mathematical Morphology method rudimentary algorithm:
The target image X given for an one and construction operator S, S is moved on the image, in each present bit
X is set, Sx there are 3 kinds of possible states:
1)That is Sx and X maximal correlation；
22)I.e. Sx is uncorrelated to X, XcFor the supplementary set of X；
3) Sx ∩ X and Sx ∩ XcIt is not sky, i.e., Sx is related to X section；
All structural elements and the maximal correlation point set of image of constituting that definition meets point x 1) are exactly corrosion of the S to X,
Be denoted as X Θ S is indicated with aggregate manner are as follows:
Expansion on the contrary, be that the every bit in X is expanded as Sx, is denoted as with corrosionIs defined as:
Enlarged image is expanded, the cavity after image segmentation in object can be filled up with the part being broken in Contiguous graphics.
Another object of the present invention is to provide a kind of Euproctis of species discrimination method for realizing the Euproctis insect
The species identification system of the species identification system of insect, the Euproctis insect includes:
Image capture module is connect with main control module, is entomologized front and rear wing by image capture module using image pick-up device
Image data information；
Main control module, with image capture module, image processing module, gene database information extraction modules, cloud service mould
Block, data memory module, display module connection, work normally for controlling modules by single-chip microcontroller；
Image processing module is connect with main control module, for being detected by image of the image processing software to acquisition
Identification；Main control module dispatches image processing module and carries out detection identification using image of the image processing software to acquisition, extracts poison
Moth wing image math-morphological features, main control module scheduling image processing module analyze math-morphological features, filter out use
In the characteristic parameter of Classification and Identification；
Poison moth wing image data base information extraction modules, connect, for downloading Euproctis by network library with main control module
Insect poison moth wing image data information simultaneously compares judgement；
Cloud service module, connect with main control module, for concentrating big data resource to the poison moth of acquisition by Cloud Server
Belong to insect poison moth wing image data information to be analyzed and processed；
Data memory module is connect with main control module, for the Euproctis insect poison moth wing by memory storage acquisition
Image data information；
Display module is connect with main control module, for the Euproctis insect poison moth wing image by display display acquisition
Data information.
Another object of the present invention is to provide a kind of Information Numbers of species discrimination method using the Euproctis insect
According to processing terminal.
Advantages of the present invention and good effect are as follows:
The present invention can automatically and precisely recognize Euproctis caste by image processing module, can solve traditional classification
Method labor intensive, time and the bad problem of efficiency；The method of the present invention is simple, easily operated, time-consuming short.
The present invention proposes a kind of error compensation method for discrete sampling, and this method only needs calibration single compensation parameter,
Suitable for all video cameras and algorithm, it can be obviously improved original image procossing precision, for centroid method and Gauss curve fitting method, the benefit
Image zooming-out precision can be increased to 0.03pixel by the method for repaying；Using the wavelet image blending algorithm based on region clarity,
By comparing the articulation index between different each neighborhood of pixels of original image wavelet transformation subgraph, the detailed information of different scale is extracted
Panorama focusedimage is constructed, to eliminate the blocking artifact phenomenon that traditional airspace fusion method is generated due to using single fusion scale.
The algorithm carries out multi-resolution decomposition using wavelet transformation to multiple original images first；Then it is directed to low frequency subgraph, with neighborhood of pixels
Interior point acutance and low frequency fusion coefficients are obtained as fusion measurement, for high frequency subgraph, according to the La Pu in neighborhood of pixels
The sum of Laplacian operater is merged, to reduce influence of the local noise for clarity operation result；Finally by small echo inversion
Acquisition blending image is changed, fusion method of the kind based on region clarity is better than other algorithms most in use；Using k-means algorithm letter
Single, fast convergence rate is expansible and high-efficient.
Main control module scheduling image processing module of the present invention carries out detection knowledge using image of the image processing software to acquisition
Not；
Appropriate thresholding is set according to figure minimum containment rectangle length-width ratio, is filtered；
Thresholding is set according to side length each in the figure of source and the minimum value of perimeter ratio, removes the surpriseization part in targeted graphical；
Abbreviation processing is made to targeted graphical number of edges, makes that there is identical number of edges with source figure；Obtain image preprocessing identification knot
Fruit；Obtain image preprocessing recognition result；
To the processing of image preprocessing recognition result enhancement；In acquisition source figure and targeted graphical eigenmatrix it is most like to
The Euclidean distance of amount and maximum phase and coefficient；Biological tissue's attenuation coefficient is rebuild again；Rebuild biological tissue scatters coefficient；
It can get accurate image information, provide guarantee for the processing of next step.
The present invention uses BP neural network classifier, and the method for building up of BP neural network classifier can be to imitate faster
Rate and more accurately classification identify poison moth.
The present invention is during image processing module extracts poison moth wing image math-morphological features, using Mathematical Morphology method
Rudimentary algorithm can make the poison moth wing image math-morphological features more micronization processes of acquisition, reduce the presence of error.
Data memory module in the present invention, by memory storage acquisition Euproctis insect poison moth wing picture number it is believed that
During breath, in order to which storage demand and Algorithms T-cbmplexity can be reduced using the overall data excavation side based on small echo
Method.
Detailed description of the invention
Fig. 1 is the species discrimination method flow chart that the present invention implements the Euproctis insect provided.
Fig. 2 is the species identification system structure chart that the present invention implements the Euproctis insect provided.
In figure: 1, image capture module；2, main control module；3, image processing module；4, gene database information extraction mould
Block；5, cloud service module；6, data memory module；7, display module.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, with reference to embodiments, to the present invention
It is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, it is not used to
Limit the present invention.
With reference to the accompanying drawing and specific embodiment is further described application principle of the invention.
As shown in Figure 1, the species discrimination method of Euproctis insect provided in an embodiment of the present invention includes:
S101 is entomologized the image data information of front and rear wing by image capture module using image pick-up device；
S102, main control module dispatch image processing module and carry out detection knowledge using image of the image processing software to acquisition
Not, poison moth wing image math-morphological features are extracted, main control module analyzes math-morphological features, filters out and knows for classifying
Another characteristic parameter；
S103 downloads Euproctis insect poison moth using network data base by poison moth wing image data base information extraction modules
Wing image data information compares judgement；And concentrate big data resource to acquisition using Cloud Server by cloud service module
Image information is handled；
S104 utilizes the poison moth wing image data information of memory storage acquisition by data memory module；And by aobvious
Show module using the poison moth wing image data information of display display acquisition.
As shown in Fig. 2, the species identification system of Euproctis insect provided by the invention includes: image capture module 1, master control
Module 2, image processing module 3, poison moth wing image data base information extraction modules 4, cloud service module 5, data memory module 6,
Display module 7.
Image capture module 1 is connect with main control module 2, before and after being entomologized by image capture module using image pick-up device
The image data information of wing；
Main control module 2 takes with image capture module 1, image processing module 3, gene database information extraction modules 4, cloud
Business module 5, data memory module 6, display module 7 connect, and control modules normal work for passing through single-chip microcontroller；
Image processing module 3 is connect with main control module 2, for being examined by image of the image processing software to acquisition
Survey identification；Main control module 2 dispatches image processing module and carries out detection identification using image of the image processing software to acquisition, extracts
Poison moth wing image math-morphological features, main control module scheduling image processing module are analyzed math-morphological features, are filtered out
Characteristic parameter for Classification and Identification；
Poison moth wing image data base information extraction modules 4, connect, for downloading poison moth by network library with main control module 2
Belong to insect poison moth wing image data information and compares judgement；
Cloud service module 5 is connect with main control module 2, for concentrating big data resource to the poison of acquisition by Cloud Server
Moth category insect poison moth wing image data information is analyzed and processed；
Data memory module 6 is connect with main control module 2, for the Euproctis insect poison moth by memory storage acquisition
Wing image data information；
Display module 7 is connect with main control module 2, for the Euproctis insect poison moth wing figure by display display acquisition
As data information.
Described image processing module 3 extracts poison moth wing image math-morphological features, dispatches image procossing mould in main control module
Block analyzes math-morphological features, filters out the characteristic parameter for Classification and Identification, during identifying to poison moth,
For faster efficiency and more accurately classification, using BP neural network classifier, the method for building up of BP neural network classifier,
The following steps are included:
Step 1, using 3 layers of BP neural network of a hidden layer, input layer number 5 is corresponding through principal component analysis
The F of foundation1、F2、F3、F4And F5Variable；
Step 2, output layer neuron number are corresponding moth category caste, and for quantitative analysis, different moths is belonged to elder brother
Worm carries out assignment；
Step 3, hidden layer neuron number are set to variable, are calculated according to following empirical equation；
In formula: InIndicate input layer number, OnIndicate output layer neuron number, a ∈ [1,10]；
Step 4, input layer to hidden layer use tansig transmission function；Hidden layer to output layer transmission function be purelin；
Training algorithm uses non-linear damped least-square method optimization algorithm, function trainlm；
Step 5, training error maximum cycle are 2000, network training result and training sample data relative error
Control is 1 × 10-6Within, it is believed that reach required precision, learning coefficient 0.05.
During described image processing module 3 extracts poison moth wing image math-morphological features, in order to make the poison moth of acquisition
Wing image math-morphological features more micronization processes, using Mathematical Morphology method rudimentary algorithm are as follows:
The target image X given for an one and construction operator S, S is moved on the image, in each present bit
X is set, Sx there are 3 kinds of possible states:
1)That is Sx and X maximal correlation；
2)I.e. Sx is uncorrelated to X, XcFor the supplementary set of X；
3) Sx ∩ X and Sx ∩ XcIt is not sky, i.e., Sx is related to X section；
All structural elements and the maximal correlation point set of image of constituting that definition meets point x 1) are exactly corrosion of the S to X,
Be denoted as X Θ S is indicated with aggregate manner are as follows:
Expansion on the contrary, be that the every bit in X is expanded as Sx, is denoted as with corrosionIs defined as:
Enlarged image is expanded, the cavity after image segmentation in object can be filled up with the part being broken in Contiguous graphics；
According to expansion and the definition of erosion operation, there are also following a few morphologic basic operations of class；
It opens:That is the result after X is corroded by S is expanded by S again；
It closes:That is the result after X is expanded by S is corroded by S again；
It hits: setting X as image collection to be studied, S is construction operator, and S is by two non-intersecting part S1And S2Composition, i.e.,
S=S1∪S2And S1∩S2=H；
X is expressed as by S hitIts result is defined as:
And
That is the result that X is hit by S is still an image, wherein every x must simultaneously meet two conditions: S1Quilt
It is included in X after x translation, and S2After being translated by x not including.
The data memory module 6 stores the Euproctis insect poison moth wing image data information of acquisition by memory
In the process, the overall data method for digging based on small echo is used in order to which storage demand and Algorithms T-cbmplexity can be reduced,
The following steps are included:
Step 1, selects an orthogonal representation, which is suitable for that the type of data model will be constructed；
Step 2 generates an approximate orthonormal basis coefficient in each local sites；
Step 3 from each station to single station, the proximate selection sample of mobile data collection, and generates corresponding non-linear friendship
Pitch the approximate base system number of item；
Step 4 synthesizes partial model, by model conversation at the form being understood that, and output model.
Below with reference to concrete analysis, the invention will be further described.
The species discrimination method of Euproctis insect provided in an embodiment of the present invention, the species identification side of the Euproctis insect
Method includes:
Main control module dispatches image processing module and carries out detection identification using image of the image processing software to acquisition；Specifically
Include:
Appropriate thresholding is set according to figure minimum containment rectangle length-width ratio, is filtered；
Thresholding is set according to side length each in the figure of source and the minimum value of perimeter ratio, removes the surpriseization part in targeted graphical；
Abbreviation processing is made to targeted graphical number of edges, makes that there is identical number of edges with source figure；Obtain image preprocessing identification knot
Fruit；Obtain image preprocessing recognition result；
To the processing of image preprocessing recognition result enhancement；In acquisition source figure and targeted graphical eigenmatrix it is most like to
The Euclidean distance of amount and maximum phase and coefficient；Biological tissue's attenuation coefficient is rebuild again；Rebuild biological tissue scatters coefficient；
Rebuilding attenuation coefficient includes:
Using the parallel of the integrated image collection device space uniform distribution of image processing module in image after the pre-treatment
Light is to sample ginIt is irradiated, measures incident intensity by acquiring the irradiation light that no sample is blocked；It is rightThe right and left is the same as divided by ginAnd take negative logarithm, then:
Collect 360 degree of measurement data G0Afterwards, inverse Radon is realized using accurate efficient filter back-projection reconstruction algorithm
Transformation calculates attenuation coefficient, i.e. μt=FBP (G0)；
Scattering coefficient is rebuild by formulag1Include image collection device
The influence scattered in imaging, as the angle acquisition data g from a certain determination1When,WithScattering angle determine, coefficient k one
The constant of a determination；Both sides are the same as divided by kgin, then have:
Known by above formulaProlong for scattering coefficientThe weighting Radon in direction is converted, institute weighted value ω1(t) and ω2(t)
It is function related with attenuation coefficient, it willDiscretization is simultaneously expressed as follows with a matrix type:
Wμs=G1；
Wherein W indicates the weight matrix after discretization, μsAnd G1Scattering coefficient vector sum different angle is respectively indicated to measure
The AVHRR NDVI vector arrived establishes following objective function using the weighted least-squares criterion with penalty function:
Wherein the first item of expression formula is the approximate expression form of likelihood function, Section 2 R (μs) it is regular terms, usual root
It is constructed according to the prior information of image, β is regularization factors, and Matrix C is covariance matrix；With niIndicate the inspection of image collection device
The scattered photon number measured, corresponding covariance matrix indicate are as follows:
Using optimal method to Φ (μs) objective function solve, that is, find out scattering coefficient:
μs=argmin Φ (μs)；
According to biological tissue's attenuation coefficient, biological tissue scatters coefficient, the absorption coefficient of sample is calculated, utilizes relational expression μt
=μa+μsCalculate the absorption coefficient μ of samplea；
The Euclidean distance of most like vector and maximum phase and coefficient are specific in acquisition source figure and targeted graphical eigenmatrix
Include:
Firstly, establishing the eigenmatrix P of source figure P and targeted graphical Q respectively counterclockwiseEAnd QE:
PE=[P1 T P2 T … P2N-1 T P2N T]；
QE=[Q1 T Q2 T … Q2N-1 T Q2N T]；
Euclidean distance formula d (x, y) and included angle cosine formula sim (x, y) are as follows:
With d (x, y) and it is the basis sim (x, y), redefines two matrix Ds and S, make:
Find out the minimum value in D and S；
Eu is enabled respectivelye=min { Dij, 1≤i≤j=2N；Sime=max { Sij, 1≤i≤j=2N；
Then the eigenmatrix of needle directional structure vectorical structure figure P and Q, the above-mentioned calculation method of repetition find out two features in order again
Minimum value Eu in matrix between most complete vectorcAnd Simc；
Finally enable Eu=min { Eue, Euc}；
Sim=min { Sime, Simc}；
Eu and Sim be two figure of P, Q correspond to most like vector Euclidean distance and it is maximum mutually and coefficient；
Include: to the processing of image preprocessing recognition result enhancement
Initial vector is carried out once to repeatedly deformation, on the basis of with adjacent corner sequence structure initial vector, then
The geometrical characteristic for adding figure, using the adjacent corner of order of addition than as new initial vector；Initial vector is carried out
Multiple nonlinear processing is once arrived, carries out evolution processing using by initial vector；
Multiple similarity calculation is carried out to deformed initial vector, finally by weighted average value, with Euclidean distance Eu
It is as follows with the evaluation formula mutually with coefficient S im:
N is the number of vector deformation, k in above formulaiFor weight coefficient, EuiAnd SimiVector is European after deforming for i-th
Distance, Eu (P, Q) are the evaluation of Euclidean distance, n=4, kiTake 0.25.
Image processing module is provided with image modification module, image modification module using high-precision error compensating method into
The amendment of row image, specifically:
If any coordinate for extracting point is (xi, yi), the standard deviation of fitting isAnd k=f1(P2) and P2=f2(σ)；
The data of step 1 are updated to following formula:
Δ x=(k-1) x, (- 0.5 < x < 0.5)；
To obtain the offset of every X-direction:
X in formulaixFor xiFractional part (- 0.5 < xix< 0.5), then obtain compensated coordinate (xi+xp, yi+yp)；
It is improved in image definition in display module, the place of image is carried out using the Image Fusion based on wavelet transformation
Reason, specifically has:
Step 1 carries out wavelet transformation to original image respectively, constructs the small echo Pyramid transform of image, obtains each straton image
Horizontal, vertical, diagonal direction matrix of wavelet coefficients；
Step 2 selects sharpness evaluation function, calculates the definition values of each location of pixels of high and low frequency wavelet subband, leads to
Fusion rule is crossed, fusion treatment is carried out to the matrix of wavelet coefficients on each decomposition layer, obtains final fusion coefficients；
Sharpness evaluation function are as follows:
In formula: seeking the absolute value of each point and its 8 neighborhood point gray scale difference weighted sum in neighborhood, then by m × n, institute in neighborhood
There is a resulting value to be added divided by total number of pixels, df is grey scale change amplitude, distance increment of the dx between pixel, the power of weighted sum
Value is related at a distance from point-to-point transmission, and the more close then weight of distance is bigger, and neighbours' point weight both horizontally and vertically is taken here
1, neighbours' point weight of diagonal takes
For high-frequency signal image take four neighborhood Laplace operator sums square as sharpness evaluation function, described four
Neighborhood Laplace operator unimodality is good, high sensitivity, the obvious pixel of variation tendency near focusing surface, four neighborhoods of f (i, j)
Laplacian operator calculation formula is as follows:
In formula:
Then high frequency imaging locates clarity at pixel (i, j) is defined as:
Fusion rule: for low frequency subgraph, the definition values at each pixel are calculated using acutance method, using adaptive
Criterion obtains the low frequency coefficient of blending image, and fusion coefficients calculating formula is as follows:
In formula: EAAnd EBRespectively indicate original image A, B, the definition values of low frequency subgraph point, HLFor customized thresholding threshold
Value, when the definition values difference of two images is more than threshold value, then the fusion coefficients at this are directed to clear area；
Otherwise fusion coefficients weight is determined by the definition values of two width subgraphs；
Step 3 carries out wavelet inverse transformation to wavelet coefficient after fusion, and resulting reconstructed image is final blending image；
It is as follows using k-means algorithm to search element for data in the main control module:
If object set M={ x1, x2..., xn, xi={ xi1, xi2..., xit), sample xiWith sample xjIt is European away from
It is as follows from calculation formula:
d(xi, xj)=[(xi1-xj1)2+(xi2-xj2)2+…+(xin-xjn)2]；
Square criterion error function is as follows:
In formula: k is the number to be clustered；tiFor the number of sample in the i-th class；niIt is the mean value of sample in the i-th class.
Described image processing module processing method further comprises:
(1) detection image of at least insect in a detection zone is obtained with a visual detector；
(2) it according to a background image of the detection zone and the detection image, is obtained about this extremely by background subtracting method
One first foreground image of a few insect；
(3) saturation degree in first foreground image is obtained, and eliminates the non-characteristic area in first foreground image
Block is to obtain one second foreground image；
(4) feature in second foreground image about an at least insect is obtained according to a threshold values；
(5) classification of an at least insect is judged according to this feature.
In the above-described embodiments, can come wholly or partly by software, hardware, firmware or any combination thereof real
It is existing.When using entirely or partly realizing in the form of a computer program product, the computer program product include one or
Multiple computer instructions.When loading on computers or executing the computer program instructions, entirely or partly generate according to
Process described in the embodiment of the present invention or function.The computer can be general purpose computer, special purpose computer, computer network
Network or other programmable devices.The computer instruction may be stored in a computer readable storage medium, or from one
Computer readable storage medium is transmitted to another computer readable storage medium, for example, the computer instruction can be from one
A web-site, computer, server or data center pass through wired (such as coaxial cable, optical fiber, Digital Subscriber Line (DSL)
Or wireless (such as infrared, wireless, microwave etc.) mode is carried out to another web-site, computer, server or data center
Transmission).The computer-readable storage medium can be any usable medium or include one that computer can access
The data storage devices such as a or multiple usable mediums integrated server, data center.The usable medium can be magnetic Jie
Matter, (such as: floppy disk, hard disk, tape), optical medium (for example, DVD) or semiconductor medium (such as solid state hard disk Solid
State Disk (SSD)) etc..
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all in essence of the invention
Made any modifications, equivalent replacements, and improvements etc., should all be included in the protection scope of the present invention within mind and principle.
Claims (5)
1. a kind of species discrimination method of Euproctis insect, which is characterized in that the species discrimination method packet of the Euproctis insect
It includes:
The first step is entomologized the image data information of front and rear wing using image pick-up device；
Second step carries out detection identification using image of the image processing software to acquisition, it is special to extract poison moth wing image Mathematical Morphology
Sign, main control module analyze math-morphological features, filter out the characteristic parameter for Classification and Identification；
Third step compares judgement using network data base downloading Euproctis insect poison moth wing image data information；And pass through
Cloud service module concentrates big data resource to handle the image information of acquisition using Cloud Server；
4th step utilizes the poison moth wing image data information of memory storage acquisition by data memory module；And pass through display
The poison moth wing image data information that module is acquired using display display.
2. the species discrimination method of Euproctis insect as described in claim 1, which is characterized in that extract poison moth wing image mathematics
Morphological feature analyzes math-morphological features, filters out the characteristic parameter for Classification and Identification, using BP neural network point
Class device, the method for building up of BP neural network classifier, comprising the following steps:
Step 1, using 3 layers of BP neural network of a hidden layer, input layer number 5, correspondence is established through principal component analysis
F1、F2、F3、F4And F5Variable；
Step 2, output layer neuron number are that different moth category insects is carried out assignment by corresponding moth category caste；
Step 3, hidden layer neuron number are set to variable, are calculated according to following empirical equation；
In formula: InIndicate input layer number, OnIndicate output layer neuron number, a ∈ [1,10]；
Step 4, input layer to hidden layer use tansig transmission function；Hidden layer to output layer transmission function be purelin；Training
Algorithm uses non-linear damped least-square method optimization algorithm, function trainlm；
Step 5, training error maximum cycle are 2000, and network training result and training sample data relative error control
1 × 10-6Within, it is believed that reach required precision, learning coefficient 0.05.
3. the species discrimination method of Euproctis insect as described in claim 1, which is characterized in that extract poison moth wing image mathematics
Morphological feature uses Mathematical Morphology method rudimentary algorithm:
The target image X given for an one and construction operator S, S is moved on the image, in each current location x,
Sx has 3 kinds of possible states:
1)That is Sx and X maximal correlation；
2)I.e. Sx is uncorrelated to X, XcFor the supplementary set of X；
3) Sx ∩ X and Sx ∩ XcIt is not sky, i.e., Sx is related to X section；
All structural elements and the maximal correlation point set of image of constituting that definition meets point x 1) are exactly corrosion of the S to X, are denoted as X
Θ S is indicated with aggregate manner are as follows:
Expansion on the contrary, be that the every bit in X is expanded as Sx, is denoted as with corrosionIs defined as:
Enlarged image is expanded, the cavity after image segmentation in object can be filled up with the part being broken in Contiguous graphics.
4. a kind of species of Euproctis insect for the species discrimination method for realizing Euproctis insect described in claim 1 identify system
System, which is characterized in that the species identification system of the Euproctis insect includes:
Image capture module is connect with main control module, is entomologized the figure of front and rear wing by image capture module using image pick-up device
As data information；
Main control module, with image capture module, image processing module, gene database information extraction modules, cloud service module, number
It connects according to memory module, display module, is worked normally for controlling modules by single-chip microcontroller；
Image processing module is connect with main control module, for carrying out detection identification by image of the image processing software to acquisition；
Main control module dispatches image processing module and carries out detection identification using image of the image processing software to acquisition, extracts poison moth wing figure
As math-morphological features, main control module scheduling image processing module analyzes math-morphological features, filters out for classifying
The characteristic parameter of identification；
Poison moth wing image data base information extraction modules, connect with main control module, for downloading Euproctis insect by network library
Poison moth wing image data information simultaneously compares judgement；
Cloud service module, connect with main control module, for concentrating big data resource to the Euproctis elder brother of acquisition by Cloud Server
Worm poison moth wing image data information is analyzed and processed；
Data memory module is connect with main control module, for the Euproctis insect poison moth wing image by memory storage acquisition
Data information；
Display module is connect with main control module, for the Euproctis insect poison moth wing image data by display display acquisition
Information.
5. a kind of information data of species discrimination method using Euproctis insect described in claims 1 to 3 any one is handled
Terminal.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811324618.8A CN109344917A (en) | 2018-11-08 | 2018-11-08 | A kind of species identification method and identification system of Moth insects |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811324618.8A CN109344917A (en) | 2018-11-08 | 2018-11-08 | A kind of species identification method and identification system of Moth insects |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109344917A true CN109344917A (en) | 2019-02-15 |
Family
ID=65314246
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811324618.8A Pending CN109344917A (en) | 2018-11-08 | 2018-11-08 | A kind of species identification method and identification system of Moth insects |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109344917A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111444984A (en) * | 2020-04-24 | 2020-07-24 | 江苏省农业科学院宿迁农科所 | Intelligent insect defining algorithm in ecological breeding system |
| CN111476180A (en) * | 2020-04-08 | 2020-07-31 | 杭州国际旅行卫生保健中心(杭州海关口岸门诊部) | Interactive automatic identification method and system for blood-sucking insects |
| CN111898680A (en) * | 2020-07-31 | 2020-11-06 | 陈艳 | A biometric identification method based on multi-view morphological images of inspection materials and deep learning |
| CN112021293A (en) * | 2020-09-03 | 2020-12-04 | 常德鑫芙蓉环保有限公司 | Intelligent tobacco warehouse fumigation pest killing method |
| CN112042618A (en) * | 2020-09-03 | 2020-12-08 | 常德鑫芙蓉环保有限公司 | Fumigating and pest killing method for tobacco |
| CN113906482A (en) * | 2019-06-03 | 2022-01-07 | 拜耳公司 | System for determining the effect of active substances on mites, insects and other organisms in test plates with cavities |
| CN116298348A (en) * | 2023-05-17 | 2023-06-23 | 深圳迈瑞动物医疗科技股份有限公司 | Blood analysis device and method for species |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101533517A (en) * | 2009-04-15 | 2009-09-16 | 北京联合大学 | Structure feature based on Chinese painting and calligraphy seal image automatic extracting method |
| CN107729534A (en) * | 2017-10-30 | 2018-02-23 | 中原工学院 | Caste identifying system and method based on big data Cloud Server |
-
2018
- 2018-11-08 CN CN201811324618.8A patent/CN109344917A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101533517A (en) * | 2009-04-15 | 2009-09-16 | 北京联合大学 | Structure feature based on Chinese painting and calligraphy seal image automatic extracting method |
| CN107729534A (en) * | 2017-10-30 | 2018-02-23 | 中原工学院 | Caste identifying system and method based on big data Cloud Server |
Non-Patent Citations (2)
| Title |
|---|
| 王和平等: "数学形态法在指纹特征提取中的研究与应用", 《中北大学学报》 * |
| 蔡小娜等: "用于昆虫分类鉴定的人工神经网络方法研究:主成分分析与数学建模", 《生物数学学报》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113906482A (en) * | 2019-06-03 | 2022-01-07 | 拜耳公司 | System for determining the effect of active substances on mites, insects and other organisms in test plates with cavities |
| CN111476180A (en) * | 2020-04-08 | 2020-07-31 | 杭州国际旅行卫生保健中心(杭州海关口岸门诊部) | Interactive automatic identification method and system for blood-sucking insects |
| CN111444984A (en) * | 2020-04-24 | 2020-07-24 | 江苏省农业科学院宿迁农科所 | Intelligent insect defining algorithm in ecological breeding system |
| CN111444984B (en) * | 2020-04-24 | 2021-05-28 | 江苏省农业科学院宿迁农科所 | Intelligent insect defining algorithm in ecological breeding system |
| CN111898680A (en) * | 2020-07-31 | 2020-11-06 | 陈艳 | A biometric identification method based on multi-view morphological images of inspection materials and deep learning |
| CN112021293A (en) * | 2020-09-03 | 2020-12-04 | 常德鑫芙蓉环保有限公司 | Intelligent tobacco warehouse fumigation pest killing method |
| CN112042618A (en) * | 2020-09-03 | 2020-12-08 | 常德鑫芙蓉环保有限公司 | Fumigating and pest killing method for tobacco |
| CN116298348A (en) * | 2023-05-17 | 2023-06-23 | 深圳迈瑞动物医疗科技股份有限公司 | Blood analysis device and method for species |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109344917A (en) | A kind of species identification method and identification system of Moth insects | |
| Mittal et al. | An efficient edge detection approach to provide better edge connectivity for image analysis | |
| Bauer et al. | The potential of automatic methods of classification to identify leaf diseases from multispectral images | |
| Zhang et al. | Saliency detection based on self-adaptive multiple feature fusion for remote sensing images | |
| Sampathkumar et al. | An automated crop and plant disease identification scheme using cognitive fuzzy C-means algorithm | |
| Kumar et al. | Review on image segmentation techniques | |
| CN108510499B (en) | A kind of image threshold segmentation method and device based on fuzzy set and Otsu | |
| Hou et al. | Infrared and visible images fusion using visual saliency and optimized spiking cortical model in non-subsampled shearlet transform domain | |
| CN108537751B (en) | Thyroid ultrasound image automatic segmentation method based on radial basis function neural network | |
| Shen et al. | Biomimetic vision for zoom object detection based on improved vertical grid number YOLO algorithm | |
| US10346980B2 (en) | System and method of processing medical images | |
| CN107437068B (en) | Pig individual identification method based on Gabor direction histogram and pig body hair pattern | |
| Hao et al. | A hierarchical region-merging algorithm for 3-D segmentation of individual trees using UAV-LiDAR point clouds | |
| Ghazal et al. | Automated framework for accurate segmentation of leaf images for plant health assessment | |
| CN109325490A (en) | THz image target recognition method based on deep learning and RPCA | |
| CN114241511A (en) | A weakly supervised pedestrian detection method, system, medium, device and processing terminal | |
| Zhao et al. | The stress detection and segmentation strategy in tea plant at canopy level | |
| Cao et al. | Case instance segmentation of small farmland based on Mask R-CNN of feature pyramid network with double attention mechanism in high resolution satellite images | |
| Wang et al. | Object-oriented change detection approach for high-resolution remote sensing images based on multiscale fusion | |
| CN109191434A (en) | Image detecting system and detection method in a kind of cell differentiation | |
| CN117649549B (en) | A pest identification method and system based on image analysis | |
| Bao et al. | Method for wheat ear counting based on frequency domain decomposition of MSVF-ISCT | |
| Nickfarjam et al. | Multi-resolution gray-level image enhancement using particle swarm optimization | |
| Siddiqui et al. | Clustering techniques for image segmentation | |
| Wu et al. | Super-resolution fusion optimization for poultry detection: A multi-object chicken detection method |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190215 |
|
| RJ01 | Rejection of invention patent application after publication |