CN105160354A - Apple disease identifying method based on sparse representation - Google Patents
Apple disease identifying method based on sparse representation Download PDFInfo
- Publication number
- CN105160354A CN105160354A CN201510508653.5A CN201510508653A CN105160354A CN 105160354 A CN105160354 A CN 105160354A CN 201510508653 A CN201510508653 A CN 201510508653A CN 105160354 A CN105160354 A CN 105160354A
- Authority
- CN
- China
- Prior art keywords
- image
- disease
- apple
- scab
- class
- 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
Classifications
-
- 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
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2136—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on sparsity criteria, e.g. with an overcomplete basis
-
- 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
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10024—Color image
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30181—Earth observation
- G06T2207/30188—Vegetation; Agriculture
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
Abstract
The present invention discloses an apple disease identifying method based on sparse representation, comprising the following steps of: S1: carrying out scab image segmentation of a diseased apple to obtain a binary scab image; S2: constructing a sparse representation optimization problem; S3, solving the optimization problem by using an iterative method so as to obtain a sparse coefficient; and S4, calculating a residual error of a training sample and an unknown disease category sample to be determined, and using a category to which the minimum residual error belongs as a disease category of the sample to be determined. The apple disease identifying method based on sparse representation, which is provided by the present invention, has the advantages of high feature extraction speed, high identification rate, stable identification effect, high real-time performance, easiness for implementation and the like.
Description
Technical field
The present invention relates to image procossing and pattern-recognition, be specifically related to the apple disease recognition methods based on rarefaction representation.
Background technology
Apple is distributed widely in the multiple area of China and even the world, is one of main fruit of eating of multinational resident, and containing abundant carbohydrates, vitamin, trace element, pectin etc. in apple, these substances on human's healths have a lot of benefit.But apple is that one is easily caught an illness fruit, the disease species of apple is a lot, and endangering larger disease has rot, early defoliation disease, Apple with Fruit-bagged Treatment diplostomiasis, rust, powdery mildew etc.Apple disease control all has consequence and effect in apple life cycle.Because numerous orchard workers' disease prevention techniques level is not high, prevent and treat extensive, it is commonplace that abuse phenomenon disorderly joined by agricultural chemicals, and some orchard workers are in order to controlling disease, germinate from apple tree, bloom, overall process all periodically sprinkle variety classes agricultural chemicals of result, cause pesticide residues in apples severe overweight thus.Science, rationally, safely prevent and treat apple disease, to healthy, the result time limit that extends fruit tree of people, increase the fruit tree life-span and reduce environmental pollution etc. all significant.
The fact shows, finding apple disease situation and correct diagnosis disease classification early, is the prerequisite of carrying out scientific prevention and cure apple disease.Only know the classification of disease, just can suit the remedy to the case.Due to the color of different diseases apple scab, widely different between texture and shape, so orchard worker is often according to the classification of apple scab diagnosis disease.Because disease apple scab is complicated, various, irregular, and disease initial stage scab feature is obvious not, so be often difficult to disease is made to science, diagnosed exactly based on the method for manual observation.Along with improving constantly of computing machine and image processing techniques, automatic plant disease recognition methods research is the more popular research direction of of academic circles at present.The apple recognition method of a lot of classics all needs the feature extracting apple scab image, but due to the complicacy of apple scab, makes the recognition effect of these methods not good.
Rarefaction representation is a kind of new target identification method, is widely applied in recognition of face.The people such as Wright propose rarefaction representation sorting technique [WringtJ, YangAY, GaneshA, etal.Robustfacerecognitionviasparserepresentation.IEEETr ansactionsonPatternAnalysisandMachineIntelligence.2009,31 (2): 210-227].The method effectively can solve feature selecting in face recognition process and to the robustness problem blocked.Although sparse representation method obtains extensive investigation and application at field of target recognition, seldom see that the research be applied in fruit disease identification is reported by sparse representation method.
Summary of the invention
Technical matters to be solved by this invention is the not good problem of existing apple disease recognition effect.
In order to solve the problems of the technologies described above, the technical solution adopted in the present invention is to provide a kind of apple disease recognition methods based on rarefaction representation, comprises the following steps:
S1, the scab Iamge Segmentation of disease apple, convert every width disease Apple image to gray level is 0 ~ 255, dimension is 32 × 32 gray level image matrix, be designated as g (x, y), x and y represents the coordinate figure of certain pixel in this image, utilizes the Probability p that the arbitrary gray level of histogram method computed image is i
i;
Appointing establishes an initial threshold d that each pixel of disease Apple image is divided into C by gray-scale value size
1and C
2two classes, calculate C
1the probability sum of class and gray average are respectively
Calculate C
2the probability sum of class and gray average are respectively
The overall average gray scale of computed image is
μ=ω
1μ
1+ω
2μ
2(3)
The inter-class variance of computed image is
σ=ω
1(μ
1-μ)
2+ω
2(μ
2-μ)
2(4)
Change d values from 1 to 255, segmentation threshold when making σ in formula (4) get maximal value, as the optimal threshold of Iamge Segmentation, is designated as D, by D segmentation disease Apple image, obtains the scab image of disease apple binaryzation;
S2, builds rarefaction representation optimization problem, supposes there is the binaryzation scab image of the known apple disease classification of n width k class as training set, in training set, if the i-th class has n
iwidth scab image, i=1,2 ..., k, then n=n
1+ n
2+ ...+n
k, the every width image in n width image is lined up a m dimensional vector by row, wherein, all n of the i-th class
iwidth image vectorization forms matrix
each row of this matrix all represent a width vectorization image, the n width vectorization image construction matrix W=[W of all k classes
1, W
2..., W
k];
According to sparse coding principle, the nonnegative matrix optimization problem of a structure regularization:
Wherein, W is basis matrix, and X is matrix of coefficients, || X||
1,1=∑
ij| X
ij|, X
ijrepresent the projection coefficient of y in W on an i-th class jth training sample, λ > 0 is regular parameter;
Solve X by alternative manner, iterative formula is as follows
X←(-X.*min(0,(λ/2)E-W
Ty))./(W
TWX)(6)
In formula (6): E to be an element be entirely 1 matrix, ' .* ' represents that the corresponding element in two matrixes is long-pending, and ' ./' represents the business of the corresponding element in two matrixes;
S3, utilize the optimization problem of solution by iterative method formula (6), solution procedure is as follows:
(1), input regular parameter λ and maximum iteration time M, need to choose λ and M according to actual classification problem, the span of λ is 0.5 ~ 0.9;
(2), the row of standardization y and W, make it have unit l
2norm, initialization non-negative n dimensional vector X;
(3), a=(λ/2) E-W is calculated
ty, E to be element be entirely 1 matrix;
(4), iteration i=1:M
x←(-x.*min(0,a))./(A
TAx);
Iteration terminates, and obtains sparse coefficient;
S4, the residual error of calculation training sample and unknown disease classification sample undetermined, for any i-th class binaryzation scab image, definition δ
ix () is n dimensional vector, wherein a δ
ix the nonzero element of () is only the x part relevant with the i-th class, calculate the residual epsilon of the i-th class training sample and unknown classification sample y
i(y)=|| y-A δ
i(x) ||
2, i=1,2 ..., k; Select residual error minimum value
affiliated classification, is the recognition result of y, as the disease classification of sample undetermined.
Above-mentioned based in the apple disease recognition methods of rarefaction representation, in S1, the function in Matlab software ' imread ' and ' rgb2gray ' is utilized to convert every width disease Apple image to gray level is 0 ~ 255, dimension is 32 × 32 gray level image matrix.
Above-mentioned based in the apple disease recognition methods of rarefaction representation, in S1, utilize the histogram of function ' histc ' computed image in Matlab software, calculating pixel is the Probability p of i
i:
In formula (7), r
ifor gray-scale value is the pixel count of i, N is total pixel number.
Above-mentioned based in the apple disease recognition methods of rarefaction representation, in S1, split disease Apple image by D,
Then, design the square structure element of 3 × 3, utilize closure operation level and smooth disease Apple image G (x, y) border fill the breach of scab inside, again the separate section of scab is linked together, finally, Glycerine enema is carried out to eliminate the noise around scab to the scab region obtained, obtains the scab image of disease apple binaryzation.
Apple disease recognition methods based on rarefaction representation provided by the invention can overcome existing apple disease recognition methods because apple scab is complicated, various, it is irregular to arrange and the color of scab, shape and texture such as to change in time at the reason problem such as cause the not high and recognition effect of disease recognition rate unstable.
Embodiment
Below in conjunction with embodiment, the present invention is further described.
Apple disease recognition methods based on rarefaction representation provided by the invention comprises the segmentation of apple scab, builds the optimization problem of rarefaction representation, asks sparse coefficient and disease classification identification four part.
The first step, scab is split.
The function in Matlab software ' imread ' and ' rgb2gray ' is utilized to convert every width disease Apple image to gray level is 0 ~ 255, dimension is 32 × 32 gray level image matrix, be designated as g (x, y), x and y represents the coordinate figure of certain pixel in this image; Utilize the histogram of function ' histc ' computed image in Matlab software; Calculating pixel is the Probability p of i
i:
In formula, r
ifor gray-scale value is the pixel count of i, N is total pixel number.
Appoint and establish an initial image segmentation threshold value d, by d, disease Apple image is divided into C
1and C
2two classes, calculate C
1the probability sum of class and gray average are respectively
Calculate C
2the probability sum of class and gray average are respectively
The overall average gray scale of computed image is
μ=ω
1μ
1+ω
2μ
2(4)
The inter-class variance of computed image is
σ=ω
1(μ
1-μ)
2+ω
2(μ
2-μ)
2(5)
Change d values from 1 to 255, calculate segmentation threshold corresponding when making σ in formula (5) get maximal value, be designated as D, using the optimal threshold of D as Iamge Segmentation, by D segmentation disease Apple image,
Then, design the square structure element of 3 × 3, utilize closure operation level and smooth disease Apple image G (x, y) border and fill the breach of scab inside, then the separate section of scab is linked together; Finally, Glycerine enema is carried out to eliminate the noise around scab to the scab region obtained, obtains the scab image of disease apple binaryzation.
Second step, builds the optimization problem of rarefaction representation.
The identification of apple disease classification is exactly utilize the sample in training set to determine the classification of unknown sample.Suppose there is the binaryzation scab image of the known apple disease classification of n width k class as training set.In training set, if the i-th class has n
iwidth scab image, i=1,2 ..., k, then n=n
1+ n
2+ ...+n
k; Every width image in n width image is lined up a m dimensional vector by row, wherein, all n of the i-th class
iwidth image vectorization forms matrix
each row of this matrix all represent a width vectorization image.The n width of all k classes forms matrix W=[W
1, W
2..., W
k].
According to sparse coding principle, the nonnegative matrix optimization problem of a structure regularization:
Wherein, W is basis matrix, and X is matrix of coefficients, || X||
1,1=∑
ij| X
ij|, X
ijrepresent the projection coefficient of y in W on an i-th class jth training sample, λ > 0 is regular parameter.
Solve X by alternative manner, iterative formula is as follows
X←(-X.*min(0,(λ/2)E-W
Ty))./(W
TWX)(8)
In formula (8): E to be an element be entirely 1 matrix, ' .* ' represents that the corresponding element in two matrixes is long-pending, and ' ./' represents the business of the corresponding element in two matrixes.
3rd step, solves.Utilize the optimization problem of solution by iterative method formula (6).
Solution procedure is as follows:
(1) input regular parameter λ and maximum iteration time M, need to choose λ and M according to actual classification problem, the span of λ is 0.5 ~ 0.9.
(2) row of standardization y and W, make it have unit l
2norm, initialization non-negative n dimensional vector X;
(3) a=(λ/2) E-W is calculated
ty, E to be element be entirely 1 matrix;
(4) iteration i=1:M
x←(-x.*min(0,a))./(A
TAx);
Iteration terminates.
4th step, the identification of apple disease classification.
For any i-th class binaryzation scab image, definition δ
ix () is n dimensional vector, wherein a δ
ix the nonzero element of () is only the x part relevant with the i-th class.Calculate the residual epsilon of the i-th class training sample and unknown classification sample y
i(y)=|| y-A δ
i(x) ||
2, i=1,2 ..., k; Select residual error minimum value
affiliated classification, is the recognition result of y.
A specific embodiment of the present invention is as follows:
The first step, scab is split.
Utilize the mono-anti-digital camera of Canon EOS700D to gather the ring spot of apple, anthracnose and each 50 width of acne pinta three kinds of disease geo-radar images in apple orchard in September-10:00-11:00 time period in the morning in October, selection every class disease geo-radar image 30 width is wherein as training sample set arbitrarily again, and each 20 width of remaining every class image are as test sample book set.First the function in Matlab software ' imread ' is utilized to convert all disease Apple images to digital picture, each digital picture is converted to the gray level image matrix that gray level is 0 ~ 255, dimension is 32 × 32 by recycling function ' rgb2gray ', any one is designated as g (x, y), x and y represents the coordinate figure of certain pixel in this image; The function in Matlab software ' histc ' is utilized to calculate the histogram of each gray level image; Calculating pixel is the Probability p of i
i:
In formula, r
ifor gray-scale value is the pixel count of i, N is total pixel number.
Appoint and establish an initial image segmentation threshold value d, by d, all pixels of g (x, y) are divided into two classes: the pixel set that pixel value is less than or equal to d is designated as C
1class, and the pixel set that pixel value is greater than d is designated as C
2two classes, calculate C
1the probability sum of class and gray average are respectively
Calculate C
2the probability sum of class and gray average are respectively
The overall average gray scale of computed image is
μ=ω
1μ
1+ω
2μ
2(4)
The inter-class variance of computed image is designated as
σ=ω
1(μ
1-μ)
2+ω
2(μ
2-μ)
2(5)
Change d values from 1 to 255, calculate segmentation threshold corresponding when making σ in formula (5) get maximal value, be designated as D, using the optimal threshold of D as Iamge Segmentation, by D segmentation disease Apple image, obtain scab image, be designated as G (x, y):
Then, design the square structure element of 3 × 3, utilize closure operation level and smooth disease Apple image G (x, y) border and fill the breach of scab inside, then the separate section of scab is linked together; Finally, Glycerine enema is carried out to eliminate the noise around scab to the scab region obtained, obtains the scab image of disease apple binaryzation.
Second step, builds the optimization problem of rarefaction representation.
The identification of apple disease classification is exactly utilize training sample to determine the classification of unknown sample.Obtain by the first step scab image splitting rear three kinds of disease Apple images, every width scab image dimension is 32 × 32.Wherein each 30 width of every class totally 90 width in training set, each 20 width of every class totally 60 width in test set.In training set, the every width image in 90 width images is lined up 32 × 32=1024 dimensional vector by row, wherein, 30 vector composition matrix W that the i-th class obtains
i=[w
i1, w
i2..., W
i30], i=1,2,3, each row of this matrix all represent a width vectorization image.Totally 90 vectors of all 3 classes form a basis matrix W=[W
1, W
2, W
3].
According to sparse coding principle, the nonnegative matrix optimization problem of a structure regularization:
In formula (7): X is matrix of coefficients, || X||
1,1=∑
ij| X
ij|, X
ijrepresent the projection coefficient of y in W on an i-th class jth training sample, λ > 0 is regular parameter.
Solve X by alternative manner, iterative formula is as follows
X←(-X.*min(0,(λ/2)E-W
Ty))./(W
TWX)(8)
In formula (8): E to be an element be entirely 1 matrix, ' .* ' represents that the corresponding element in two matrixes is long-pending, and ' ./' represents the business of the corresponding element in two matrixes.
3rd step, solves.Utilize the optimization problem of iteration French (8) model solution formula (7).
Solution procedure is as follows:
(5) regular parameter λ=0.5 and maximum iteration time M=50 is inputted.
(6) row of standardization y and W, make it have unit l
2norm, initialization non-negative 90 dimensional vector X=[1,1 ..., 1];
(7) a=(λ/2) E-W is calculated
ty, E to be element be entirely 1 matrix;
(8) iteration i=1:M
X←(-X.*min(0,a))./(W
TWx);
Iteration terminates.
4th step, the identification of apple disease classification.
For any i-th class binaryzation scab image, definition δ
ix () is n dimensional vector, wherein a δ
ix the nonzero element of () is only the x part relevant with the i-th class.Calculate the residual error δ of the i-th class training sample and unknown classification sample y
i(y)=|| y-A δ
i(x) ||
2, i=1,2 ..., k; Select residual error minimum value
affiliated classification, is the recognition result of y.
Repeat identifying above 50 times, calculate the average recognition result of 50 experiments and the average recognition result of three kinds of apple diseases, 50 experiments.The recognition result of algorithm is given in table 1.
Disease blade | Discrimination |
Ring spot | 94.24 |
Anthracnose | 94.63 |
Acne pinta | 96.17 |
Average recognition rate | 95.01 |
The recognition result of table 1 three kinds of apple disease blades
The present invention is to provide a kind of apple disease recognition methods represented based on the scab image sparse of disease apple, according to obtained rarefaction representation coefficient, calculate the residual error of unknown classification sample and all kinds of training sample, determine the classification of unknown classification sample according to residual error.Method provided by the invention does not need to extract the characteristic of division of Apple image, the advantage such as have that discrimination is high, recognition effect stable and practical.
At present, the present invention realizes on common PC computing machine, to operating system not requirement.Be mainly used in computer based apple disease classification recognition system, may be used for the apple disease identification in actual production.
The present invention is not limited to above-mentioned preferred forms, and anyone should learn the structure change made under enlightenment of the present invention, and every have identical or close technical scheme with the present invention, all falls within protection scope of the present invention.
Claims (4)
1., based on the apple disease recognition methods of rarefaction representation, it is characterized in that, comprise the following steps:
S1, the scab Iamge Segmentation of disease apple, convert every width disease Apple image to gray level is 0 ~ 255, dimension is 32 × 32 gray level image matrix, be designated as g (x, y), x and y represents the coordinate figure of certain pixel in this image, utilizes the Probability p that the arbitrary gray level of histogram method computed image is i
i;
Appointing establishes an initial threshold d that each pixel of disease Apple image is divided into C by gray-scale value size
1and C
2two classes, calculate C
1the probability sum of class and gray average are respectively
Calculate C
2the probability sum of class and gray average are respectively
The overall average gray scale of computed image is
μ=ω
1μ
1+ω
2μ
2(3)
The inter-class variance of computed image is
σ=ω
1(μ
1-μ)
2+ω
2(μ
2-μ)
2(4)
Change d values from 1 to 255, segmentation threshold when making σ in formula (4) get maximal value, as the optimal threshold of Iamge Segmentation, is designated as D, by D segmentation disease Apple image, obtains the scab image of disease apple binaryzation;
S2, builds rarefaction representation optimization problem, supposes there is the binaryzation scab image of the known apple disease classification of n width k class as training set, in training set, if the i-th class has n
iwidth scab image, i=1,2 ..., k, then n=n
1+ n
2+ ...+n
k, the every width image in n width image is lined up a m dimensional vector by row, wherein, all n of the i-th class
iwidth image vectorization forms matrix
each row of this matrix all represent a width vectorization image, the n width vectorization image construction matrix W=[W of all k classes
1, W
2..., W
k];
According to sparse coding principle, the nonnegative matrix optimization problem of a structure regularization:
Wherein, W is basis matrix, and X is matrix of coefficients, || X||
1,1=∑
ij| X
ij|, X
ijrepresent the projection coefficient of y in W on an i-th class jth training sample, λ > 0 is regular parameter;
Solve X by alternative manner, iterative formula is as follows
X←(-X.*min(0,(λ/2)E-W
Ty))./(W
TWX)(6)
In formula (6): E to be an element be entirely 1 matrix, ' .* ' represents that the corresponding element in two matrixes is long-pending, and ' ./' represents the business of the corresponding element in two matrixes;
S3, utilize the optimization problem of solution by iterative method formula (6), solution procedure is as follows:
(1), input regular parameter λ and maximum iteration time M, need to choose λ and M according to actual classification problem, the span of λ is 0.5 ~ 0.9;
(2), the row of standardization y and W, make it have unit l
2norm, initialization non-negative n dimensional vector X;
(3), a=(λ/2) E-W is calculated
ty, E to be element be entirely 1 matrix;
(4), iteration i=1:M
x←(-x.*min(0,a))./(A
TAx);
Iteration terminates, and obtains sparse coefficient;
S4, the residual error of calculation training sample and unknown disease classification sample undetermined, for any i-th class binaryzation scab image, definition δ
ix () is n dimensional vector, wherein a δ
ix the nonzero element of () is only the x part relevant with the i-th class, calculate the residual epsilon of the i-th class training sample and unknown classification sample y
i(y)=|| y-A δ
i(x) ||
2, i=1,2 ..., k; Select residual error minimum value
affiliated classification, is the recognition result of y, as the disease classification of sample undetermined.
2. as claimed in claim 1 based on the apple disease recognition methods of rarefaction representation, it is characterized in that, in S1, the function in Matlab software ' imread ' and ' rgb2gray ' is utilized to convert every width disease Apple image to gray level is 0 ~ 255, dimension is 32 × 32 gray level image matrix.
3. as claimed in claim 1 based on the apple disease recognition methods of rarefaction representation, it is characterized in that, in S1, utilize the histogram of function ' histc ' computed image in Matlab software, calculating pixel is the Probability p of i
i:
In formula (7), r
ifor gray-scale value is the pixel count of i, N is total pixel number.
4. as claimed in claim 1 based on the apple disease recognition methods of rarefaction representation, it is characterized in that, in S1, split disease Apple image by D,
Then, design the square structure element of 3 × 3, utilize closure operation level and smooth disease Apple image G (x, y) border fill the breach of scab inside, again the separate section of scab is linked together, finally, Glycerine enema is carried out to eliminate the noise around scab to the scab region obtained, obtains the scab image of disease apple binaryzation.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510508653.5A CN105160354A (en) | 2015-08-19 | 2015-08-19 | Apple disease identifying method based on sparse representation |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510508653.5A CN105160354A (en) | 2015-08-19 | 2015-08-19 | Apple disease identifying method based on sparse representation |
Publications (1)
Publication Number | Publication Date |
---|---|
CN105160354A true CN105160354A (en) | 2015-12-16 |
Family
ID=54801206
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510508653.5A Pending CN105160354A (en) | 2015-08-19 | 2015-08-19 | Apple disease identifying method based on sparse representation |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105160354A (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107292874A (en) * | 2017-06-29 | 2017-10-24 | 深圳前海弘稼科技有限公司 | The control method and device of crop disease |
CN108319927A (en) * | 2018-02-13 | 2018-07-24 | 南京工程学院 | A kind of method of automatic identification disease |
CN108960172A (en) * | 2018-07-12 | 2018-12-07 | 南京工程学院 | A method of identification GPR image Damage Types |
CN112749675A (en) * | 2021-01-20 | 2021-05-04 | 云南农业大学 | Potato disease identification method based on convolutional neural network |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102819748A (en) * | 2012-07-19 | 2012-12-12 | 河南工业大学 | Classification and identification method and classification and identification device of sparse representations of destructive insects |
CN103530658A (en) * | 2013-10-15 | 2014-01-22 | 武汉科技大学 | Method for recognizing plant leaf data based on sparse representation |
CN103593652A (en) * | 2013-10-28 | 2014-02-19 | 西京学院 | Cucumber disease identification method based on cucumber leaf symptom image processing |
-
2015
- 2015-08-19 CN CN201510508653.5A patent/CN105160354A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102819748A (en) * | 2012-07-19 | 2012-12-12 | 河南工业大学 | Classification and identification method and classification and identification device of sparse representations of destructive insects |
CN103530658A (en) * | 2013-10-15 | 2014-01-22 | 武汉科技大学 | Method for recognizing plant leaf data based on sparse representation |
CN103593652A (en) * | 2013-10-28 | 2014-02-19 | 西京学院 | Cucumber disease identification method based on cucumber leaf symptom image processing |
Non-Patent Citations (6)
Title |
---|
IAN ALEXANDER MERWIN ETAL: "Long-term orchard groundcover management systems affect soil microbial communities and apple replant disease", 《REGULAR ARTICLE》 * |
JIE TIAN ETAL: "Karhunen-Loeve Transform and Sparse Representation Based Plant Leaf Disease Recognition", 《TELKOMNIKA INDONESIAN JOURNAL OF ELECTRICAL ENGINEERING》 * |
毕傲睿: "苹果叶子病害图像识别系统的设计与实现", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
毛亮 等: "基于稀疏场水平集的荔枝图像分割算法", 《农业工程学报》 * |
胡秋霞: "基于图像分析的植物叶部病害识别方法研究", 《中国博士学位论文全文数据库 信息科技辑》 * |
霍迎秋 等: "基于压缩感知理论的苹果病害识别方法", 《农业机械学报》 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107292874A (en) * | 2017-06-29 | 2017-10-24 | 深圳前海弘稼科技有限公司 | The control method and device of crop disease |
CN108319927A (en) * | 2018-02-13 | 2018-07-24 | 南京工程学院 | A kind of method of automatic identification disease |
CN108319927B (en) * | 2018-02-13 | 2021-11-02 | 南京工程学院 | Method for automatically identifying diseases |
CN108960172A (en) * | 2018-07-12 | 2018-12-07 | 南京工程学院 | A method of identification GPR image Damage Types |
CN108960172B (en) * | 2018-07-12 | 2021-12-07 | 南京工程学院 | Method for identifying GPR image disease type |
CN112749675A (en) * | 2021-01-20 | 2021-05-04 | 云南农业大学 | Potato disease identification method based on convolutional neural network |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Adeel et al. | Diagnosis and recognition of grape leaf diseases: An automated system based on a novel saliency approach and canonical correlation analysis based multiple features fusion | |
Suh et al. | Transfer learning for the classification of sugar beet and volunteer potato under field conditions | |
Jhuria et al. | Image processing for smart farming: Detection of disease and fruit grading | |
Hung et al. | Orchard fruit segmentation using multi-spectral feature learning | |
Tang et al. | A multi-stage framework with context information fusion structure for skin lesion segmentation | |
CN108388896A (en) | A kind of licence plate recognition method based on dynamic time sequence convolutional neural networks | |
CN105574534A (en) | Significant object detection method based on sparse subspace clustering and low-order expression | |
CN107229904A (en) | A kind of object detection and recognition method based on deep learning | |
JP6341650B2 (en) | Image processing apparatus, image processing method, and program | |
CN105678284A (en) | Fixed-position human behavior analysis method | |
CN110533683B (en) | Image omics analysis method fusing traditional features and depth features | |
CN109190559A (en) | A kind of gesture identification method, gesture identifying device and electronic equipment | |
CN102663757A (en) | Semi-automatic image cutting method based on nuclear transfer | |
CN105160354A (en) | Apple disease identifying method based on sparse representation | |
CN106570874A (en) | Image marking method combining local image constraint and overall target constraint | |
Wang et al. | Tea picking point detection and location based on Mask-RCNN | |
CN107239532A (en) | Data digging method and device | |
CN108334878A (en) | Video images detection method and apparatus | |
CN108596195A (en) | A kind of scene recognition method based on sparse coding feature extraction | |
Hou et al. | Detection and localization of citrus fruit based on improved You Only Look Once v5s and binocular vision in the orchard | |
Suparyanto et al. | Detecting Hemileia vastatrix using vision AI as supporting to food security for smallholder coffee commodities | |
Defriani et al. | Recognition of Regional Traditional House in Indonesia Using Convolutional Neural Network (CNN) Method | |
Miao et al. | Crop weed identification system based on convolutional neural network | |
WO2024078112A1 (en) | Method for intelligent recognition of ship outfitting items, and computer device | |
CN111160372B (en) | Large target identification method based on high-speed convolutional neural network |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20151216 |
|
WD01 | Invention patent application deemed withdrawn after publication |