CN110147802A - The inertinite classification method and system of trend fluction analysis are gone based on multi-fractal - Google Patents
The inertinite classification method and system of trend fluction analysis are gone based on multi-fractal Download PDFInfo
- Publication number
- CN110147802A CN110147802A CN201910392730.3A CN201910392730A CN110147802A CN 110147802 A CN110147802 A CN 110147802A CN 201910392730 A CN201910392730 A CN 201910392730A CN 110147802 A CN110147802 A CN 110147802A
- Authority
- CN
- China
- Prior art keywords
- inertinite
- fractal
- image
- trend
- classification
- 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
- G06F18/24—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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/40—Analysis of texture
- G06T7/41—Analysis of texture based on statistical description of texture
- G06T7/42—Analysis of texture based on statistical description of texture using transform domain methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/40—Analysis of texture
- G06T7/41—Analysis of texture based on statistical description of texture
- G06T7/48—Analysis of texture based on statistical description of texture using fractals
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/42—Global feature extraction by analysis of the whole pattern, e.g. using frequency domain transformations or autocorrelation
- G06V10/431—Frequency domain transformation; Autocorrelation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Probability & Statistics with Applications (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses inertinite classification methods and system that trend fluction analysis is removed based on multi-fractal, belong to technical field of image processing, comprising the following steps: S1: acquisition inertinite sample；S2: inertinite textural characteristics are extracted；S3: classify to inertinite.In the step S1, HD type microphotometer is selected to be acquired the optical imagery of coal petrography inertinite, in the step S2, in a computer using collected inertinite micro-image as the storage of the two-dimensional matrix of gray level image, trend fluction analysis method is gone to be analyzed using multi-fractal.The present invention can effectively improve inertinite each component discrimination, facilitate and distinguish different classes of coal petrography micro-image, be conducive to assess and the coal petrography purifying research work efficiently utilized coal petrography technological property.
Description
Technical field
The present invention relates to technical field of image processing, and in particular to the inertinite of trend fluction analysis is removed based on multi-fractal
Classification method and system.
Background technique
As people are continuously increased clean energy resource and natural resources demand, study the scheme that its high-efficiency cleaning utilizes and draw
Play extensive concern.Coal separation is to utilize the basis of Filter Tuber For Clean Coal and premise and the most economical effective method of clean coal technology.It realizes
The automatic classification and identification of macerals, the purifying efficient utilization to the effective assessment and coal petrography of coal petrography technological property
With important research significance.
Inertinite is the microstructural important component of coal petrography, the classification and identification of inertinite each component is realized, to coal
Effective assessment of rock technological property and purifying efficient utilize of coal petrography have important research significance.Inertinite texture structure
Clearly, textural characteristics are relatively stable and in coalification course, therefore texture probes into each maceral classification method of inertinite
Desired characteristics.
In recent years, with the development of fractal theory, classification someone of coal petrography is taken and introduces fractal box and wavelet packet
Energy square this single fractal method describes its textural characteristics.However coal petrography inertinite texture structure is complicated, multiplicity, using list
One fractal theory is difficult to that complicated structure is described, and also can not effectively portray the irregular characteristics of image in part, and
Multi-fractal Theory in point shape is based on multiple dimensioned, multirange material surface character description method, is portraying object
There is good effect in singularity.
Summary of the invention
Technical problem to be solved by the present invention lies in: the optical image recognition rate of coal petrography inertinite how is improved, is provided
Go based on multi-fractal the inertinite classification method of trend fluction analysis.
The present invention be by the following technical programs solution above-mentioned technical problem, the present invention the following steps are included:
S1: acquisition inertinite sample
The optical imagery of coal petrography inertinite is acquired；
S2: inertinite textural characteristics are extracted
One-dimensional MF-DFA (multi-fractal detrend fluctuation analysis) method is introduced into higher-dimension multiple analysis, two dimension is obtained
Multi-fractal goes trend fluction analysis method, goes trend fluction analysis to extract inertinite micro-image using two-dimentional multi-fractal
Textural characteristics, obtain descriptor Hurst index α (q) and singular spectrum f (α), formula is as follows:
α (q)=τ ' (q)=h (q)+qh ' (q)
F (α)=q α (q)-τ (q)=q [α-h (q)]+2
Wherein, α (q) is used to describe the local singular intensity of texture image, and calculating is combined by the collection with identical α (q) value
At the fractal dimension of union obtain f (α), singular spectrum f (α) is used to describe the unusual intensity of the overall situation of texture image；
S3: classify to inertinite
Support vector machines RBF (the diameter based on radial base is constructed using radial basis function as kernel function using ballot method
To basic function)-SVM (support vector machines) classifier, classify to inertinite.
Preferably, in the step S1, HD type microphotometer is selected to carry out the optical imagery of coal petrography inertinite
Acquisition.
Preferably, in the step S2, using collected inertinite micro-image as the two-dimensional matrix of gray level image
Storage in a computer, goes trend fluction analysis algorithm to be analyzed using multi-fractal.
Preferably, the analytic process in the step S2 specifically includes the following steps:
S201: M is divided by plane is nonoverlapping using the square area that size is s × ss×Ns (Ms≡[M/s],
Ns≡ [N/s]), each subregion is denoted as Xm,n=Xm,n(i, j), wherein Xm,n(i, j)=X (r+i, t+j), 1≤i, j≤s, r
=(m-1) s, t=(n-1) s；
S202: calculate each sub-regions accumulation and, calculation formula is as follows:
Wherein, 1≤i, j≤s, middle m=1,2 ..., Ms, n=1,2 ..., Ns；
S203: it is fitted by bivariate polynomial of order one and obtains Gm,nLocal trendFitting formula is as follows:
Wherein, 1≤i, j≤s, a, b, c are the free parameters that can be determined by least square method；
S204: the trend of two-dimensional matrix is eliminated by accumulation and with the difference of previous step fitting function, obtains residual error sequence
Arrange ym,n(i, j), residual sequence formula are as follows:
S205: X is calculatedm,nRemove trend wave function F (m, n, s), go trend wave function formula as follows:
S206: calculating q rank wave function, and formula is as follows:
Change the size for dividing rectangle side length s, usual s minimum takes 6, is up to (M, N)/4, repeats step S207；
S207: if Fq(s) there is long-range power rate dependence with s variation, relevance formula is as follows:
Fq(s)∝sh(q)
Then calculate lnFq(s) to the linear regression of s, scaling exponent h (q) number, i.e. generalized Hurst index are obtained；
S208: the q value different to each, according to standard partition function, available corresponding performance figure τ (q),
Formula is as follows:
τ (q)=qh (q)-Df
Wherein, DfIt is the fractal dimension of Multifractal Method geometry supported collection；
S209: broad sense Multifractal Dimension D is calculatedq, formula is as follows:
S210: the descriptor Hurst index α (q) of the important unusual intensity of characterization of two dimensional gray surface another two is calculated
With singular spectrum f (α).
Preferably, in the step S202, Gm,n=Gm,n(i, j) (i, j=1,2 .., s) itself is a plane.
Preferably, in the step S208, micro-image is estimated for two dimension, selects Df=2.
Preferably, in the step S3, vote method the following steps are included:
S301: several inertinites image of all categories is chosen respectively, as test sample；
S302: the texture characteristic amount of test sample image is calculated；
S303: it is sequentially inputted to the characteristic quantity being calculated as test sample data in each RBF-SVM classifier；
S304: counting the number of votes obtained of each classification, and the number of votes obtained of highest classification is determined as to the classification of input picture, realizes
Classification to each classification of inertinite.
A kind of inertinite categorizing system for removing trend fluction analysis based on multi-fractal, comprising:
Sampling module is acquired for the optical imagery to coal petrography inertinite；
Characteristic extracting module, the texture for going trend fluction analysis to extract inertinite micro-image by multi-fractal are special
Sign；
Categorization module, for classifying using ballot method and in conjunction with RBF-SVM classifier to inertinite；
Central processing module, for sending instruction into relevant action to modules；
The sampling module, characteristic extracting module, categorization module are electrically connected with central processing module.
The present invention has the advantage that this removes the inertinite point of trend fluction analysis based on multi-fractal compared with prior art
Class method, first acquisition coal petrography inertinite sample, then go trend fluction analysis to extract inertinite micrograph using multi-fractal
The textural characteristics of picture；Ballot method is finally used, classifies in conjunction with RBF kernel function support vector machine, inertinite can be effectively improved
Each component discrimination facilitates and distinguishes different classes of coal petrography micro-image, is conducive to for the assessment of coal petrography technological property and coal petrography
The purifying research work efficiently utilized.
Detailed description of the invention
Fig. 1 is workflow schematic diagram of the invention；
Fig. 2 is typical coal petrography inertinite micro-image in the embodiment of the present invention；
Fig. 3 is the multifractal spectra of fire-burning fusinite in the embodiment of the present invention；
Fig. 4 is the multifractal spectra of degradofusinite in the embodiment of the present invention；
Fig. 5 is the multifractal spectra of semifusinite in the embodiment of the present invention；
Fig. 6 is the multifractal spectra of resene body in the embodiment of the present invention；
Fig. 7 is the multifractal spectra of sclerotinite in the embodiment of the present invention；
Fig. 8 is the multifractal spectra of macrinite in the embodiment of the present invention；
Fig. 9 is the multifractal spectra of inertodetrinite in the embodiment of the present invention；
Figure 10 is the multifractal spectra of microsome in the embodiment of the present invention；
Figure 11 is the flow diagram of the ballot method class test scheme in the embodiment of the present invention.
Specific embodiment
It elaborates below to the embodiment of the present invention, the present embodiment carries out under the premise of the technical scheme of the present invention
Implement, the detailed implementation method and specific operation process are given, but protection scope of the present invention is not limited to following implementation
Example.
As shown in Figure 1, the present embodiment provides a kind of technical solutions: removing the inertinite of trend fluction analysis based on multi-fractal
Classification method, comprising the following steps:
S1: acquisition inertinite sample
As shown in Fig. 2, sampling coal petrography inertinite, is made into mating plate, HD type microscope is used under oil immersion reflected light
Photometer is acquired the image of coal petrography inertinite optical texture, and obtaining the macerals of 8 classifications, (its moderate heat burns silk quality
Body and degradofusinite are taken as 2 classification processing).
S2: inertinite textural characteristics are extracted
Store the inertinite micro-image of 8 classifications as the two-dimensional matrix of gray level image in a computer, and according to
Multi-fractal goes trend fluction analysis algorithm to be analyzed.Here, it is contemplated that in cutting procedure, image size M and N be not total
Be the multiple for dividing small square s, do not utilized so as to cause image rightmost and bottom region, we will from it is remaining its
His three directions repeat to divide.Through past trend process, the wave function of inertinite each component texture image is obtained.Therefrom may be used
To find out, in the fitting result of different rank q, trend wave function F is removedq(s) there are apparent power rates between scale s
Scaling, this shows that inertinite maceral image has self-similarity；
For different q values, by the value of each maceral h (q) of slop estimation inertinite of straight line in wave function, and
Corresponding function τ (q) is calculated by standard partition function, formula is as follows:
τ (q)=qh (q)-Df；
According to the relationship that inertinite each maceral performance figure τ (q) and q is calculated of standard partition function.When -6
When≤q≤6, it has been found that each maceral performance figure τ (q) relative to q be it is nonlinear, this can confirm inertinite
Micro-image has multi-fractal features；
By calculating, evaluation rule is obtained, formula is as follows:
α (q)=τ ' (q)=h (q)+qh ' (q)
F (α)=q α (q)-τ (q)=q [α-h (q)]+2
As shown in figs. 3-10, multiple Fractal Singular spectrum f (α) of the multifractal spectra of respectively each maceral of inertinite is one
Kind is used to portray the chain index spectrum of multi-fractal set, it provides the complete statistics to fractal inside inconsistency
Description.
S3: classify to inertinite
Using ballot method, classify in conjunction with RBF-SVM classifier to inertinite, specifically include: is aobvious to 8 kinds of inertinite
Micro-group point, every kind of 60 Zhang great little of component selection are 227 × 227 micro-image samples, carry out multi-fractal to each image and go
Trend fluction analysis obtains multifractal spectra and extracts the minimum of alpha of singular indexminValue, the maximum α of singular indexmaxValue and
Evaluation rule maximum value max (f)；We describe component using these three features and construct a three-dimensional parameter space,
Each point represents a sample in the space.
In order to realize the multi-component Classification and Identification of inertinite, the present embodiment will be using ballot classification schemes, by 8 class inertinites
Maceral combination of two constructs 28 RBF-SVM classifiers, and using grid optimizing to penalty factor c and RBF core letter
γ parameter in number optimizes.It is 227 × 227 micro-image conducts that each maceral of inertinite, which takes 60 Zhang great little, in experiment
Sample, wherein training sample each component takes 40 width, 20 width of test sample each component；28 SVM classifier effects are different,
It is each to discriminate between the classifier of two kinds of different inertinite components, as shown in table 1: the classification of each classifier is described in table
Object, and each classifier parameters of parameter for passing through grid optimizing.
The effect table of each classifier of table 1
Classifier is trained first, steps are as follows:
Inertinite 40 width of image of all categories, input sample are taken respectively；
Combining any two categories in 8 classifications is one group, totally 28 groups, then extracts each group of corresponding number of training
According to.Such as the characteristic quantity of two class component of fire-burning fusinite and degradofusinite is input in RBF-SVM1 and is trained；
All groups of training sample is input in corresponding RBF-SVM, when the training is completed, obtains 28 RBF-SVM
Classifier, these classifiers can be to 8 category classifications；
As shown in figure 11, the testing scheme of ballot method classification, steps are as follows:
S301: remaining inertinite 20 width of image of all categories is chosen respectively, as test sample；
S302: the texture characteristic amount of test sample image is calculated；
S303: it is sequentially inputted to the characteristic quantity being calculated as test sample data in 28 RBF-SVM classifiers；
S304: counting the number of votes obtained of each classification, and the number of votes obtained of highest classification is determined as to the classification of input picture, realizes
Classification to 8 classifications of inertinite.
By method of voting, the classification recognition result table of inertinite coal petrography micro-image test group is drawn.It can be concluded that each group
The recognition effect divided is preferable, and wherein degradofusinite, semifusinite, macrinite and inertodetrinite discrimination reach
100%.All components totality discrimination is 94.375%, this shows that inertinite micro-image is mentioned through past trend fluction analysis
The multi-fractal features got can effectively describe the textural characteristics of image, be to discriminate between the important information of different classes of image.
In order to further verify the superiority that multi-fractal removes trend fluction analysis in inertinite image classification, to being extracted image
Gray level co-occurrence matrixes statistic second moment, entropy, the moment of inertia and correlation use identical classification as training set and test set
Method, obtained classification results are undesirable, and sclerotinite classification accuracy rate only has 45%, general classification error rate up to 15.625%,
It is higher by 10% compared with multi-fractal features classification error rate, this is far from satisfying classificating requirement.
The present embodiment additionally provides a kind of inertinite categorizing system that trend fluction analysis is removed based on multi-fractal, comprising:
Sampling module is acquired for the optical imagery to coal petrography inertinite；
Characteristic extracting module, the texture for going trend fluction analysis to extract inertinite micro-image by multi-fractal are special
Sign；
Categorization module, for classifying using ballot method and in conjunction with RBF-SVM classifier to inertinite；
Central processing module, for sending instruction into relevant action to modules；
The sampling module, characteristic extracting module, categorization module are electrically connected with central processing module.
In conclusion the inertinite classification method and system that trend fluction analysis is removed based on multi-fractal of the present embodiment,
Coal petrography inertinite sample is acquired first, and the line of trend fluction analysis extraction inertinite micro-image is then removed using multi-fractal
Manage feature；Ballot method is finally used, classifies in conjunction with RBF kernel function support vector machine, inertinite each component can be effectively improved
Discrimination facilitates and distinguishes different classes of coal petrography micro-image, is conducive to assess coal petrography technological property and coal petrography is purifying
The research work efficiently utilized is worth being used more widely.
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 (8)
1. going the inertinite classification method of trend fluction analysis based on multi-fractal, which comprises the following steps:
S1: acquisition inertinite sample
The optical imagery of coal petrography inertinite is acquired；
S2: inertinite textural characteristics are extracted
One-dimensional MF-DFA method is introduced into higher-dimension multiple analysis, two-dimentional multi-fractal is obtained and goes trend fluction analysis method, benefit
It goes trend fluction analysis to extract the textural characteristics of inertinite micro-image with two-dimentional multi-fractal, obtains descriptor Hurst index α
(q) with singular spectrum f (α), formula is as follows:
α (q)=τ ' (q)=h (q)+qh ' (q)
F (α)=q α (q)-τ (q)=q [α-h (q)]+2
Wherein, α (q) is used to describe the local singular intensity of texture image, what calculating was combined by the collection with identical α (q) value
The fractal dimension of union obtains f (α), and singular spectrum f (α) is used to describe the unusual intensity of the overall situation of texture image；
S3: classify to inertinite
Using ballot method, classify in conjunction with RBF-SVM classifier to inertinite.
2. the inertinite classification method according to claim 1 for being removed trend fluction analysis based on multi-fractal, feature are existed
In in the step S1, selection HD type microphotometer is acquired the optical imagery of coal petrography inertinite.
3. the inertinite classification method according to claim 1 for being removed trend fluction analysis based on multi-fractal, feature are existed
In: in the step S2, computer is stored in using collected inertinite micro-image as the two-dimensional matrix of gray level image
In, go trend fluction analysis method to be analyzed using multi-fractal.
4. the inertinite classification method according to claim 1 for being removed trend fluction analysis based on multi-fractal, feature are existed
In: analytic process in the step S2 specifically includes the following steps:
S201: M is divided by plane is nonoverlapping using the square area that size is s × ss×Ns(Ms≡[M/s],Ns≡[N/
S]), each subregion is denoted as Xm,n=Xm,n(i, j), wherein Xm,n(i, j)=X (r+i, t+j), 1≤i, j≤s, r=(m-1) s,
T=(n-1) s；
S202: calculate each sub-regions accumulation and, calculation formula is as follows:
Wherein, 1≤i, j≤s, middle m=1,2 ..., Ms, n=1,2 ..., Ns；
S203: it is fitted by bivariate polynomial of order one and obtains Gm,nLocal trendFitting function is as follows:
Wherein, 1≤i, j≤s, a, b, c are the free parameters that can be determined by least square method；
S204: the trend of two-dimensional matrix is eliminated by accumulation and with the difference of previous step fitting function, obtains residual sequence ym,n
(i, j), residual sequence formula are as follows:
S205: X is calculatedm,nRemove trend wave function F (m, n, s), go trend wave function formula as follows:
S206: calculating q rank wave function, and formula is as follows:
Change the size for dividing rectangle side length s, usual s minimum takes 6, is up to (M, N)/4, repeats step S207；
S207: if Fq(s) there is long-range power rate dependence with s variation, relevance formula is as follows:
Fq(s)∝sh(q)
Then calculate lnFq(s) to the linear regression of s, scaling exponent h (q) number, i.e. generalized Hurst index are obtained；
S208: the q value different to each, according to standard partition function, available corresponding performance figure τ (q), formula
It is as follows:
τ (q)=qh (q)-Df
Wherein, DfIt is the fractal dimension of Multifractal Method geometry supported collection；
S209: broad sense Multifractal Dimension D is calculatedq, formula is as follows:
S210: the descriptor Hurst index α (q) and surprise of the important unusual intensity of characterization of two dimensional gray surface another two is calculated
Different spectrum f (α).
5. the inertinite classification method according to claim 4 for being removed trend fluction analysis based on multi-fractal, feature are existed
In: in the step S202, Gm,n=Gm,n(i, j) (i, j=1,2 .., s) itself is a plane.
6. the inertinite classification method according to claim 4 for being removed trend fluction analysis based on multi-fractal, feature are existed
In: in the step S208, micro-image is estimated for two dimension, selects Df=2.
7. the inertinite classification method according to claim 4 for being removed trend fluction analysis based on multi-fractal, feature are existed
In: in the step S3, ballot method the following steps are included:
S301: several inertinites image of all categories is chosen respectively, as test sample；
S302: the texture characteristic amount of test sample image is calculated；
S303: it is sequentially inputted to the characteristic quantity being calculated as test sample data in each RBF-SVM classifier；
S304: counting the number of votes obtained of each classification, and the number of votes obtained of highest classification is determined as to the classification of input picture, realizes to lazy
The classification of each classification of matter group.
8. a kind of inertinite categorizing system for removing trend fluction analysis based on multi-fractal characterized by comprising
Sampling module is acquired for the optical imagery to coal petrography inertinite；
Characteristic extracting module, for removing the textural characteristics of trend fluction analysis extraction inertinite micro-image by multi-fractal；
Categorization module, for classifying using ballot method and in conjunction with RBF-SVM classifier to inertinite；
Central processing module completes relevant action for sending instruction to modules；
The sampling module, characteristic extracting module, categorization module are electrically connected with central processing module.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910392730.3A CN110147802A (en) | 2019-05-13 | 2019-05-13 | The inertinite classification method and system of trend fluction analysis are gone based on multi-fractal |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910392730.3A CN110147802A (en) | 2019-05-13 | 2019-05-13 | The inertinite classification method and system of trend fluction analysis are gone based on multi-fractal |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110147802A true CN110147802A (en) | 2019-08-20 |
Family
ID=67595238
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910392730.3A Pending CN110147802A (en) | 2019-05-13 | 2019-05-13 | The inertinite classification method and system of trend fluction analysis are gone based on multi-fractal |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110147802A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110781450A (en) * | 2019-09-17 | 2020-02-11 | 广西电网有限责任公司电力科学研究院 | Bad data detection system and method for distribution feeder voltage measurement data |
| CN110954779A (en) * | 2019-11-29 | 2020-04-03 | 国网上海市电力公司 | Voltage sag source feature identification method based on S transformation and multidimensional fractal |
| CN111784738A (en) * | 2020-06-19 | 2020-10-16 | 中国科学院国家空间科学中心 | Extremely dark and weak moving target correlation detection method based on fluctuation analysis |
| CN113344859A (en) * | 2021-05-17 | 2021-09-03 | 武汉大学 | Method for quantifying capillary surrounding degree of gastric mucosa staining amplification imaging |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104808255A (en) * | 2015-04-30 | 2015-07-29 | 武汉光谷北斗控股集团有限公司 | Fractal theory-based mineralization anomaly information mining method |
| CN107609578A (en) * | 2017-08-25 | 2018-01-19 | 核工业北京地质研究院 | Remote sensing image texture analysis method based on multi-scale wavelet decomposition and fractal theory |
| CN108498106A (en) * | 2018-02-08 | 2018-09-07 | 陕西师范大学 | CNV brain electricity lie detecting methods based on multi-fractal detrend fluctuation analysis |
| CN108629364A (en) * | 2018-03-15 | 2018-10-09 | 中国石油大学(华东) | Non-gaussian type submarine geomorphy kind identification method based on multi-fractal spectrum signature |
-
2019
- 2019-05-13 CN CN201910392730.3A patent/CN110147802A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104808255A (en) * | 2015-04-30 | 2015-07-29 | 武汉光谷北斗控股集团有限公司 | Fractal theory-based mineralization anomaly information mining method |
| CN107609578A (en) * | 2017-08-25 | 2018-01-19 | 核工业北京地质研究院 | Remote sensing image texture analysis method based on multi-scale wavelet decomposition and fractal theory |
| CN108498106A (en) * | 2018-02-08 | 2018-09-07 | 陕西师范大学 | CNV brain electricity lie detecting methods based on multi-fractal detrend fluctuation analysis |
| CN108629364A (en) * | 2018-03-15 | 2018-10-09 | 中国石油大学(华东) | Non-gaussian type submarine geomorphy kind identification method based on multi-fractal spectrum signature |
Non-Patent Citations (3)
| Title |
|---|
| MARIUSZ MLYNARCZUK 等: "The application of artificial intelligence for the identification of the maceral groups and mineral components of coal", 《COMPUTERS & GEOSCIENCES》 * |
| 殷子睆: "特征抽取算法及其在煤岩显微组分分类中的应用", 《中国优秀硕士学位论文全文数据库 工程科技Ⅰ辑》 * |
| 王访: "作物诊断的叶片图像多重分形方法与建模——以油菜氮素营养和玉米病害诊断为例", 《中国博士学位论文全文数据库 基础科学辑》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110781450A (en) * | 2019-09-17 | 2020-02-11 | 广西电网有限责任公司电力科学研究院 | Bad data detection system and method for distribution feeder voltage measurement data |
| CN110781450B (en) * | 2019-09-17 | 2022-12-16 | 广西电网有限责任公司电力科学研究院 | Bad data detection method and system for distribution feeder voltage measurement data |
| CN110954779A (en) * | 2019-11-29 | 2020-04-03 | 国网上海市电力公司 | Voltage sag source feature identification method based on S transformation and multidimensional fractal |
| CN111784738A (en) * | 2020-06-19 | 2020-10-16 | 中国科学院国家空间科学中心 | Extremely dark and weak moving target correlation detection method based on fluctuation analysis |
| CN111784738B (en) * | 2020-06-19 | 2023-10-31 | 中国科学院国家空间科学中心 | Extremely dark and weak moving target association detection method based on fluctuation analysis |
| CN113344859A (en) * | 2021-05-17 | 2021-09-03 | 武汉大学 | Method for quantifying capillary surrounding degree of gastric mucosa staining amplification imaging |
| CN113344859B (en) * | 2021-05-17 | 2022-04-26 | 武汉大学 | Method for quantifying capillary surrounding degree of gastric mucosa staining amplification imaging |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110147802A (en) | The inertinite classification method and system of trend fluction analysis are gone based on multi-fractal | |
| CN109447977B (en) | Visual defect detection method based on multispectral deep convolutional neural network | |
| Kivinen et al. | Learning multiscale representations of natural scenes using Dirichlet processes | |
| RU2016122051A (en) | METHOD AND DEVICE FOR RECOGNIZING IMAGE OBJECT CATEGORY | |
| Atupelage et al. | Computational grading of hepatocellular carcinoma using multifractal feature description | |
| CN104487843A (en) | Image processing device, program, image processing method, computer-readable medium, and image processing system | |
| JP6897749B2 (en) | Learning methods, learning systems, and learning programs | |
| CN102982343B (en) | Handwritten number recognition and incremental type obscure support vector machine method | |
| CN104616000A (en) | Human face recognition method and apparatus | |
| CN103853724A (en) | Multimedia data sorting method and device | |
| CN105718937A (en) | Multi-class object classification method and system | |
| CN101996328A (en) | Wood identification method | |
| CN110781333A (en) | Method for processing unstructured monitoring data of cable-stayed bridge based on machine learning | |
| CN103886319A (en) | Intelligent held board recognizing method based on machine vision | |
| Chanda et al. | Script identification–a han and roman script perspective | |
| Sahu et al. | Image mining: a new approach for data mining based on texture | |
| CN103106412A (en) | Flake-type medium identification method and identification device | |
| Tian et al. | Support vector machine with mixture of kernels for image classification | |
| Kavitha et al. | A robust script identification system for historical Indian document images | |
| Tripathi et al. | Bag of Visual Words (BoVW) with Deep Features--Patch Classification Model for Limited Dataset of Breast Tumours | |
| Mehri et al. | Performance evaluation and benchmarking of six texture-based feature sets for segmenting historical documents | |
| Manimekalai et al. | Taxonomic classification of Plant species using support vector machine | |
| Luszczkiewicz-Piatek et al. | Effective color image retrieval based on the Gaussian mixture model | |
| Siddiqui et al. | Feature extraction for cursive language document images: Using discrete cosine transform, discrete wavelet transform and gabor filter | |
| CN103488997A (en) | Method for selecting hyperspectral image bands based on extraction of all kinds of important bands |
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 |