CN107992889A - A kind of iron spectrogram based on D-S evidence theory is as Multi-information acquisition method - Google Patents
A kind of iron spectrogram based on D-S evidence theory is as Multi-information acquisition method Download PDFInfo
- Publication number
- CN107992889A CN107992889A CN201711230790.2A CN201711230790A CN107992889A CN 107992889 A CN107992889 A CN 107992889A CN 201711230790 A CN201711230790 A CN 201711230790A CN 107992889 A CN107992889 A CN 107992889A
- Authority
- CN
- China
- Prior art keywords
- msub
- mrow
- image
- iron
- information
- 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
- G06F18/25—Fusion techniques
-
- 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
- G06F18/25—Fusion techniques
- G06F18/254—Fusion techniques of classification results, e.g. of results related to same input data
-
- 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
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- 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
- G06V10/56—Extraction of image or video features relating to colour
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/257—Belief theory, e.g. Dempster-Shafer
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Medicines Containing Antibodies Or Antigens For Use As Internal Diagnostic Agents (AREA)
Abstract
The invention discloses a kind of iron spectrogram based on D S evidence theories as Multi-information acquisition method.For iron spectrogram as the relatively low problem of heterogeneous information comprehensive utilization ratio in wear Particles Recognition, the filtering of the bianry image after Debris Image binary conversion treatment and Algorithm for Color Image Filtering are realized；By the morphological feature and color characteristic of synchronous extraction Debris Image, the hard output of support vector machines is converted into probability output using sigmoid functions, construction probability distribution function realizes that iron spectrogram is merged as the heterogeneous information of morphological feature and color characteristic.The present invention carries out Multi-information acquisition using D S evidence theories, combines advantage of the morphological feature to cutting, oxide abrasive grain sensitivity and color characteristic to slip abrasive particle sensitivity, realizes effective differentiation of three kinds of failure abrasive particles.Compared with color characteristic and morphological feature is used alone, its recognition accuracy improves more than 16.6%, effectively increases the comprehensive utilization ratio of abrasive particle information, and a kind of new thinking is provided for equipment attrition monitoring.
Description
Technical field
The invention belongs to mechanical fault diagnosis technical field, and in particular to a kind of iron spectrum based on D-S evidence theory
Image Multi-information acquisition method.
Background technology
Spectral Analysis Technology is a kind of wear particle analytical technology that International Tribology the 1970s field occurs.Number
Word image processing techniques is originated from the 1950s, having arrived rapid development between decades.The beneficial combination of the two, further
Promote the development of Spectral Analysis Technology.By analysis and discussion Debris Image, equipment attrition information can be obtained, and to the mill of equipment
Damage type and serious wear degree are prejudged.Therefore, the iron spectrum Debris Analysis based on image processing techniques receives widely
Concern and research.
Debris Image process field is composed in iron at present, two-value mainly is carried out to Debris Image using binary processing method
Change segmentation, abrasive particle is separated from background, and then feature extraction and pattern-recognition are carried out to abrasive particle.At this iron spectrogram picture
Though reason method is simple and practicable, finer filter process is lacked after binaryzation, causes to mix in the bianry image after segmentation
Enter more impurity information, be unfavorable for follow-up abrasive particle feature extraction and wear Particles Recognition.The colouring information of abrasive particle is colored iron
Spectrogram can only be ground as most basic and most direct information, oxide abrasive grain and non-ferrous metal abrasive particle by color with other types
Grain distinguishes.And the colouring information of abrasive particle lost completely during above-mentioned binary conversion treatment.
A kind of advanced information processing method that information fusion technology proposes in recent years, information fusion technology are to study how to add
Work, collaboration utilize multi-source information, and are complementary to one another various forms of information, with acquisition to the more objective of same thing or target
See, the informix treatment technology of more essential understanding.Its more succinct than the information directly obtained from each information source, less redundancy,
More useful way.Since information fusion is that same target or event are confirmed using multiple sensors or more category informations,
The confidence level of result can be improved, reduces the uncertainty of target or event.D-S evidence theory algorithm is simple, and in evidence theory
Basic probability assignment function calculate and the Evidence Combination Methods derivation of equation suffers from reliable mathematical theory basis, it is uncertain in processing
Message context has certain advantage, therefore becomes the important means of information fusion in recent years.
The content of the invention
The purpose of the invention is to improve the information from objective pattern of iron spectrogram picture and the comprehensive letter of color characteristic information
Cease utilization rate, there is provided a kind of iron spectrogram based on D-S evidence theory is as Multi-information acquisition method.This method utilizes D-S evidences
Theory, merges abrasive particle information from objective pattern and color characteristic information, accurate to improve the identification of different faults abrasive particle with this
True rate.
To reach above-mentioned purpose, the present invention is realized using following technical proposals：
A kind of iron spectrogram based on D-S evidence theory comprises the following steps as Multi-information acquisition method：
1) iron spectrum Debris Image is obtained from mechanical equipment lubrication system；
2) the iron spectrum Debris Image obtained to step 1) carries out binary conversion treatment and two-value filtering process；
3) pseudo-colour filtering processing is carried out to the iron spectrum Debris Image after step 2) processing；
4) mill is composed to the iron spectrum Debris Image after step 2) two-value filtering process and the iron after the processing of step 3) pseudo-colour filtering
Grain image extracts its information from objective pattern and color characteristic information respectively；
5) SVM points are input to as input feature vector to the information from objective pattern of extraction in step 4) and color characteristic information
0-1 is simultaneously exported and is converted into the soft output of probability by class device firmly using sigmoid functions；
6) using the soft output result of the abrasive particle morphological feature and the probability of color characteristic that are obtained in step 5) as fusion evidence
Body carries out Fusion Features, obtains recognition result.
Further improve of the invention is that the iron spectrum Debris Image obtained in step 1) includes：Slide Debris Image, cut
Cut Debris Image and oxide abrasive grain image.
Further improve of the invention is that two-value filtering process includes in step 2)：Coloured image binary conversion treatment, two
Value image abrasive particle mark, filter out foreign matter and filter out inner void.
Further improve of the invention is that colored filtering process includes in step 3)：The RGB for extracting coloured image is three
Pigment component, Algorithm for Color Image Filtering and matrix restructuring.
Further improve of the invention is that the abrasive particle morphological feature of extraction includes in step 4)：Abrasive particle area percentage,
Axial ratio, circularity and rectangular degree, the abrasive particle color characteristic extracted include：R, the three-component average of G, B and standard deviation, bag
Include 4 groups of morphological features, 6 groups of color characteristics, totally 10 groups of features.
Further improve of the invention is, is mapped to the output f (x) of SVM using sigmoid functions in step 5)
In [0,1], the expression formula of Sigmoid functions is：
Wherein, f (x) represents the output result of standard SVM；Its soft output is represented as a result, classifying just
True probability；A and B is tried to achieve by solving parameter set minimal negative log-likelihood.
Further improve of the invention is that the D-S evidence theory fusion formula in step 6) is：
Wherein, A ≠ Φ, A1,A2,…AnRepresent the subproposition of proposition A, (1-k)-1It is referred to as normalization factor, k is classics
The expression formula of conflict coefficient, wherein k is：
The fusion rule of D-S evidence theory, by above-mentioned fusion formula can by the independent information of some separate sources into
Row comprehensive utilization, improves the comprehensive utilization ratio of multi-source information.
Compared with prior art, the invention has the advantages that：
Traditional fluid graphical analysis can not extract the shape information and colouring information of abrasive particle, the information of Debris Image comprehensively
Utilization rate is low, causes fault recognition rate relatively low.The present invention carries out bianry image filtering and Algorithm for Color Image Filtering to Debris Image, point
Indescribably take its morphological feature and color characteristic；The probability output of wear Particles Recognition result is realized using SVM, and D-S cards are constructed with this
According to the BPA needed for theoretical fusion, information fusion is carried out；Compared with color characteristic and morphological feature is used alone, it identifies accurate
Rate improves more than 16.6%.
Specifically, the present invention can be obtained by being composed to iron after Debris Image carries out binary conversion treatment and two-value filtering process
Obtain information from objective pattern and retain more complete bianry image；By that can be obtained after carrying out pseudo-colour filtering processing to Debris Image
Color characteristic information retains more complete coloured image；At the iron spectrum Debris Image after two-value filtering process and pseudo-colour filtering
The morphological feature and color characteristic of iron spectrum Debris Image after reason, its morphological feature and color characteristic information extraction are more complete；
The output result of SVM classifier is exported firmly by 0-1 using sigmoid functions and is converted into the soft output of probability, information can be used as
The probability function of fusion；Finally, the present invention is melted the probability results progress feature that SVM classifier exports using D-S evidence theory
Close, the recognition result of acquisition combines the information from objective pattern and color characteristic information of Debris Image, therefore the knowledge with higher
Other accuracy rate.
Brief description of the drawings
Fig. 1 is the flow chart that the present invention realizes the filtering of Debris Image bianry image and Algorithm for Color Image Filtering；
Fig. 2 is that industry spot of the present invention gathers the process that iron composes Debris Image, and wherein a, b, c, d and e is respectively abrasive particle figure
As five committed steps of collection；
Fig. 3 is the four kinds of more typical Debris Images chosen in the present invention, and wherein a, b, c and d are respectively to choose difference
The iron spectrogram picture of color background and different colours abrasive particle；
Fig. 4 is that the process of two value filtering of Debris Image of the present invention illustrates；
Fig. 5 is filtered front and rear each component map when being Algorithm for Color Image Filtering of the present invention, wherein a, b and c is respectively before filtering
R, G, B component, d, e and) be respectively filter after R, G, B component；
Fig. 6 is result of the Debris Image of the present invention after pseudo-colour filtering；
Fig. 7 is abrasive particle morphological feature and color characteristic Fusion Model of the invention based on D-S evidence theory；
Fig. 8 is the probability output result that morphological feature is used alone in the present invention；
Fig. 9 is the recognition result that morphological feature is used alone in the present invention；
Figure 10 is the probability output result that color characteristic is used alone in the present invention；
Figure 11 is the recognition result that color characteristic is used alone in the present invention；
Figure 12 is after the present invention is merged the morphological feature of iron spectrum Debris Image and color characteristic using D-S evidence theory
Probability output result；
Figure 13 is after the present invention is merged the morphological feature of iron spectrum Debris Image and color characteristic using D-S evidence theory
Recognition result.
Embodiment
The invention will be further described with reference to the accompanying drawings and examples.
The present invention carries out more fine filtering on the basis of two value filterings first, obtains two-value filtering image.Tie at the same time
Colored iron spectrogram is closed as R, G, B three-component, obtains the coloured image for filtering out lose interest in abrasive particle and impurity.Then respectively to filtering
The bianry image of gained and coloured image carry out feature extraction afterwards, obtain the information from objective pattern and color characteristic letter of iron spectrogram picture
Breath.By adjusting the parameter setting of support vector machines (SVM), realize the probability output of abrasive particle specimen discerning result, form D-S cards
According to the probability distribution function (BPA) needed for theoretical fusion heterogeneous information.Using D-S evidence theory, to abrasive particle shape information and face
Color information is merged.Fusion results show that method proposed in this paper combines the advantage of two class different characteristic information, for three
The recognition accuracy of kind different types of faults abrasive particle has larger lifting.
Comprise the following steps that：(being related to Debris Image process part to be discussed by taking Fig. 3 a as an example)
1) iron spectrum Debris Image is obtained after some row operations such as industry spot collection lubricating oil, heated, vibration, dilution；
2) binary conversion treatment is carried out to Fig. 3 a, obtains binary image as shown in fig. 4 a, mesh in Fig. 3 a is not difficult to find out in observation
Mark abrasive particle is preferably retained；
3) bianry image abrasive particle marks, and red circle show a small amount of fine particle and impurity in Fig. 4 a, in Fig. 4 a
Each abrasive particle is marked, and calculates each abrasive particle size using eight neighborhood algorithm, abrasive particle size is calculated using MATLAB programs；
Wherein eight neighborhood calculation formula is：
4) foreign matter is filtered out.According to each abrasive particle size in bianry image, suitable threshold value T is set, Retention area is maximum
Target abrasive particle, uninterested small abrasive particle and impurity are filtered out, i.e.,：
Its filtered external result is as shown in Figure 4 b；
5) inner void is filtered out.Bianry image is negated on the basis of step 4), repeat step 3) and step 4), you can filter
Except said minuscule hole inside abrasive particle.The bianry image for filtering out foreign matter and inner void is finally obtained, as shown in figure 4d, is obtained at the same time
Obtain the two values matrix C of the image；
6) each component of Fig. 3 a coloured images is extracted.Red, blue, green three pigment component, three color of phase are extracted to Color Debris Image
Prime component is respectively as shown in Fig. 5 a, Fig. 5 b, Fig. 5 c.Three-component M × N matrix is obtained, is respectively：R=[rmn], R=[gmn], R
=[bmn].Wherein, the row and column of m and n difference representing matrix element
7) Algorithm for Color Image Filtering.Bianry image obtained by step 4) is converted into two values matrix C, using Matrix C respectively with
Matrix R, G, B corresponding element are multiplied, and realize Algorithm for Color Image Filtering.Matrix multiple calculating formula is as follows：
Wherein Rf, Gf, BfThe respectively matrix of the filtered rear gained of R, G, B；
8) matrix combines.The three-component matrix of gained after filtering is combined, construction three-dimensional array and drawing image, i.e.,
It can obtain filtered coloured image as shown in Figure 6 a；
9) heterogeneous information extraction feature.Utilize the morphological feature of the bianry image extraction iron spectrum abrasive particle obtained after filtering, bag
Include：Abrasive particle area percentage, axial ratio, circularity and rectangular degree.Utilize the coloured image extraction iron spectrum mill obtained after filtering
The color characteristic of grain, including coloured image average, standard deviation.
10) outputs firmly of the 0-1 in SVM classifier are converted into the soft output of probability using sigmoid functions, and then be converted into
The evidence body of D-S evidence theory.Sigmoid function expressions are：
Wherein, f (x) represents the output result of standard SVM；Its soft output is represented as a result, classifying just
True probability；A and B can be tried to achieve by solving parameter set minimal negative log-likelihood.
11) recognition result contrasts.Before and after verification D-S evidence theory fusion, the difference of failure wear Particles Recognition accuracy rate.Will
Recognition result after fusion is contrasted with the morphological feature to failure abrasive particle and the recognition result of color characteristic respectively, and verification is originally
The practical situations of invention.
Fluid sample standard deviation is gathered from industrial equipment scene lubricating oil, and iron spectrogram picture is obtained using iron spectrum image capturing system,
Its processing step is as shown in Figure 2.Wherein, the operation such as need to heat it, vibrate after industrial lubricant is obtained, it is ensured that mill
Grain is evenly distributed in fluid, and lubricating oil is diluted using tetrachloro-ethylene, is prevented that lubricating oil viscosity is excessive, is being prepared
Mobility is poor when iron composes spectral slice.
Fig. 1 show the filtering process flow of iron spectrum Debris Image, and the specific expansion of its flow is introduced referring particularly to above-mentioned
Step 2) is to step 8).
The present invention is using the morphological feature of iron spectrum Debris Image and color characteristic as input feature vector, to realize that foreign peoples believes
Breath fusion, Fusion Model are as shown in Figure 7.First, choose and slide abrasive particle, cutting wear particles and each 20 groups of oxide abrasive grain, repeat
Step 2) extracts the morphological feature and color characteristic of failure Debris Image to step 8) respectively.Using SVM classifier respectively to mill
Grain morphological feature and color characteristic carry out pattern-recognition.For every class failure abrasive particle, 10 groups of fault samples are chosen respectively and are instructed
Practice, remaining 10 groups of samples are used to test.Meanwhile to ensure the reliability of fusion results, for training and the morphological feature tested
And color characteristic corresponds.
Before and after verification D-S evidence theory fusion, the change of failure wear Particles Recognition accuracy rate.It is special to the form of failure abrasive particle
Sign and color characteristic carry out pattern-recognition respectively.After 4 groups of morphological feature input SVM, probability output result such as Fig. 8 institutes of gained
Show.It is as shown in Figure 9 to morphological feature recognition result can to obtain SVM as final recognition result for the maximum corresponding abrasive particle of identification probability.
Similarly, probability output result and recognition result of the SVM to 6 groups of color characteristics of abrasive particle are can obtain, as shown in Figure 10 and Figure 11.
Using D-S evidence theory composition rule, the soft output result of morphological feature and color characteristic to abrasive particle carries out letter
Breath fusion.Fusion results are as shown in figure 12, its final recognition accuracy, as shown in figure 13.
The present invention composes the problems such as information utilization existing for image processing process is relatively low for iron, proposes that one kind is based on D-S
The iron spectrogram of evidence theory is as Multi-information acquisition method.First, two-value filter is carried out on the basis of iron spectrogram is as binarization segmentation
Ripple, and extract R, G, B three-component of colored iron spectrogram, with reference to bianry image filter result, realizes the pseudo-colour filtering of iron spectrogram picture；
Secondly, by taking the image pattern of actual acquisition as an example, morphological feature is extracted to filtered bianry image respectively, to colored after filtering
Image zooming-out color characteristic；D-S information fusions finally are used as to the probability output result of each sample by the use of support vector machines
Basic probability assignment function, realizes that iron spectrogram is merged as the heterogeneous information of morphological feature and color characteristic.Fusion results show, base
In D-S evidence theory iron spectrogram as Multi-information acquisition method combine morphological feature to cutting, oxide abrasive grain it is sensitive and
Color characteristic realizes effective differentiation of three kinds of failure abrasive particles to the advantage of slip abrasive particle sensitivity.With color characteristic is used alone
Compared with morphological feature, its recognition accuracy improves more than 16.6%.
Claims (7)
1. a kind of iron spectrogram based on D-S evidence theory is as Multi-information acquisition method, it is characterised in that comprises the following steps：
1) iron spectrum Debris Image is obtained from mechanical equipment lubrication system；
2) the iron spectrum Debris Image obtained to step 1) carries out binary conversion treatment and two-value filtering process；
3) pseudo-colour filtering processing is carried out to the iron spectrum Debris Image after step 2) processing；
4) abrasive particle figure is composed to the iron spectrum Debris Image after step 2) two-value filtering process and the iron after the processing of step 3) pseudo-colour filtering
As extracting its information from objective pattern and color characteristic information respectively；
5) as input feature vector, SVM classifier is input to the information from objective pattern of extraction and color characteristic information in step 4)
And 0-1 is exported firmly using sigmoid functions and is converted into the soft output of probability；
6) using the soft output result of the abrasive particle morphological feature and the probability of color characteristic that obtain in step 5) as merge evidence body into
Row Fusion Features, obtain recognition result.
2. a kind of iron spectrogram based on D-S evidence theory according to claim 1 exists as Multi-information acquisition method, its feature
In the iron spectrum Debris Image obtained in step 1) includes：Slide Debris Image, cutting wear particles image and oxide abrasive grain image.
3. a kind of iron spectrogram based on D-S evidence theory according to claim 1 exists as Multi-information acquisition method, its feature
In two-value filtering process includes in step 2)：Coloured image binary conversion treatment, bianry image abrasive particle mark, filter out foreign matter
With filter out inner void.
4. a kind of iron spectrogram based on D-S evidence theory according to claim 1 exists as Multi-information acquisition method, its feature
In colored filtering process includes in step 3)：The RGB for extracting coloured image is three pigment components, Algorithm for Color Image Filtering and square
Battle array restructuring.
5. a kind of iron spectrogram based on D-S evidence theory according to claim 1 exists as Multi-information acquisition method, its feature
In the abrasive particle morphological feature of extraction includes in step 4)：Abrasive particle area percentage, axial ratio, circularity and rectangular degree, are carried
The abrasive particle color characteristic taken includes：R, the three-component average of G, B and standard deviation, including 4 groups of morphological features, 6 groups of color characteristics, altogether
10 groups of features.
6. a kind of iron spectrogram based on D-S evidence theory according to claim 1 exists as Multi-information acquisition method, its feature
In step 5) is middle to be mapped to the output f (x) of SVM in [0,1] using sigmoid functions, and the expression formula of Sigmoid functions is：
<mrow>
<mi>P</mi>
<mrow>
<mo>(</mo>
<mi>y</mi>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mrow>
<mi>f</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mrow>
<mn>1</mn>
<mo>+</mo>
<mi>exp</mi>
<mrow>
<mo>(</mo>
<mi>A</mi>
<mi>f</mi>
<mo>(</mo>
<mi>x</mi>
<mo>)</mo>
</mrow>
<mo>+</mo>
<mi>B</mi>
<mo>)</mo>
</mrow>
</mfrac>
</mrow>
Wherein, f (x) represents the output result of standard SVM；Its soft output is represented as a result, i.e. classification is correctly general
Rate；A and B is tried to achieve by solving parameter set minimal negative log-likelihood.
7. a kind of iron spectrogram based on D-S evidence theory according to claim 6 exists as Multi-information acquisition method, its feature
In the D-S evidence theory fusion formula in step 6) is：
<mfenced open = "" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<mi>m</mi>
<mrow>
<mo>(</mo>
<mi>A</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<msub>
<mi>m</mi>
<mn>1</mn>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>A</mi>
<mn>1</mn>
</msub>
<mo>)</mo>
</mrow>
<mo>&CirclePlus;</mo>
<msub>
<mi>m</mi>
<mn>2</mn>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>A</mi>
<mn>2</mn>
</msub>
<mo>)</mo>
</mrow>
<mo>...</mo>
<mo>&CirclePlus;</mo>
<msub>
<mi>m</mi>
<mi>n</mi>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>A</mi>
<mi>n</mi>
</msub>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mrow>
<mn>1</mn>
<mo>-</mo>
<mi>k</mi>
</mrow>
</mfrac>
<munder>
<mi>&Sigma;</mi>
<mrow>
<msub>
<mi>A</mi>
<mn>1</mn>
</msub>
<mo>&cap;</mo>
<mn>...</mn>
<msub>
<mi>A</mi>
<mi>n</mi>
</msub>
<mo>=</mo>
<mi>A</mi>
</mrow>
</munder>
<msub>
<mi>m</mi>
<mn>1</mn>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>A</mi>
<mn>1</mn>
</msub>
<mo>)</mo>
</mrow>
<mo>&times;</mo>
<msub>
<mi>m</mi>
<mn>2</mn>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>A</mi>
<mn>2</mn>
</msub>
<mo>)</mo>
</mrow>
<mo>...</mo>
<msub>
<mi>m</mi>
<mi>n</mi>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>A</mi>
<mi>n</mi>
</msub>
<mo>)</mo>
</mrow>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
Wherein, A ≠ Φ, A1,A2,…AnRepresent the subproposition of proposition A, (1-k)-1It is referred to as normalization factor, k conflicts to be classical
The expression formula of coefficient, wherein k is：
<mrow>
<mi>k</mi>
<mo>=</mo>
<munder>
<mo>&Sigma;</mo>
<mrow>
<msub>
<mi>A</mi>
<mn>1</mn>
</msub>
<mo>&cap;</mo>
<msub>
<mi>A</mi>
<mn>2</mn>
</msub>
<mo>...</mo>
<msub>
<mi>A</mi>
<mi>n</mi>
</msub>
<mo>=</mo>
<mi>&Phi;</mi>
</mrow>
</munder>
<msub>
<mi>m</mi>
<mn>1</mn>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>A</mi>
<mn>1</mn>
</msub>
<mo>)</mo>
</mrow>
<mo>&times;</mo>
<msub>
<mi>m</mi>
<mn>2</mn>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>A</mi>
<mn>2</mn>
</msub>
<mo>)</mo>
</mrow>
<mo>...</mo>
<mo>&times;</mo>
<msub>
<mi>m</mi>
<mi>n</mi>
</msub>
<mrow>
<mo>(</mo>
<msub>
<mi>A</mi>
<mi>n</mi>
</msub>
<mo>)</mo>
</mrow>
</mrow>
The fusion rule of D-S evidence theory, can be carried out the independent information of some separate sources by above-mentioned fusion formula comprehensive
Close and utilize, improve the comprehensive utilization ratio of multi-source information.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711230790.2A CN107992889A (en) | 2017-11-29 | 2017-11-29 | A kind of iron spectrogram based on D-S evidence theory is as Multi-information acquisition method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711230790.2A CN107992889A (en) | 2017-11-29 | 2017-11-29 | A kind of iron spectrogram based on D-S evidence theory is as Multi-information acquisition method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107992889A true CN107992889A (en) | 2018-05-04 |
Family
ID=62034269
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711230790.2A Pending CN107992889A (en) | 2017-11-29 | 2017-11-29 | A kind of iron spectrogram based on D-S evidence theory is as Multi-information acquisition method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107992889A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108764312A (en) * | 2018-05-17 | 2018-11-06 | 河海大学 | Optimize multi objective dam defect image detecting method based on DS |
| CN110208124A (en) * | 2019-05-30 | 2019-09-06 | 新疆大学 | The development approach of mechanical wear system based on Abrasive Wear Mechanism |
| CN110675397A (en) * | 2019-10-14 | 2020-01-10 | 国网山东省电力公司泰安供电公司 | Transformer substation protection pressing plate state checking method |
| CN112381140A (en) * | 2020-11-13 | 2021-02-19 | 国家能源集团泰州发电有限公司 | Abrasive particle image machine learning identification method based on new characteristic parameters |
| CN115457351A (en) * | 2022-07-22 | 2022-12-09 | 中国人民解放军战略支援部队航天工程大学 | Multi-source information fusion uncertainty judgment method |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103207070A (en) * | 2013-04-09 | 2013-07-17 | 苏州经贸职业技术学院 | Method for diagnosing failure of rotary machine by fusing oil and vibration |
-
2017
- 2017-11-29 CN CN201711230790.2A patent/CN107992889A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103207070A (en) * | 2013-04-09 | 2013-07-17 | 苏州经贸职业技术学院 | Method for diagnosing failure of rotary machine by fusing oil and vibration |
Non-Patent Citations (5)
| Title |
|---|
| ZHONGYU HUANG等: "《A Fault Diagnosis Method of Rolling Bearing through Wear Particle and Vibration Analyses》", 《APPLIED MECHANICS AND MATERIALS VOLS》 * |
| 曹一波等: "《基于D-S证据理论和集成神经网络的磨粒识别》", 《润滑与密封》 * |
| 权文等: "《基于SVM 概率输出与证据理论的多分类方法》", 《计算机工程》 * |
| 王永坚等: "《神经网络和证据理论融合的尾轴承磨损故障诊断》", 《中国航海》 * |
| 蒋先刚: "《基于稀疏表达的火焰与烟雾探测方法研究》", 31 August 2017, 西南交通大学出版社 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108764312A (en) * | 2018-05-17 | 2018-11-06 | 河海大学 | Optimize multi objective dam defect image detecting method based on DS |
| CN108764312B (en) * | 2018-05-17 | 2019-04-05 | 河海大学 | Optimize multi objective dam defect image detecting method based on DS |
| CN110208124A (en) * | 2019-05-30 | 2019-09-06 | 新疆大学 | The development approach of mechanical wear system based on Abrasive Wear Mechanism |
| CN110675397A (en) * | 2019-10-14 | 2020-01-10 | 国网山东省电力公司泰安供电公司 | Transformer substation protection pressing plate state checking method |
| CN110675397B (en) * | 2019-10-14 | 2023-03-28 | 国网山东省电力公司泰安供电公司 | Transformer substation protection pressing plate state checking method |
| CN112381140A (en) * | 2020-11-13 | 2021-02-19 | 国家能源集团泰州发电有限公司 | Abrasive particle image machine learning identification method based on new characteristic parameters |
| CN112381140B (en) * | 2020-11-13 | 2024-02-06 | 国家能源集团泰州发电有限公司 | Abrasive particle image machine learning identification method based on new characteristic parameters |
| CN115457351A (en) * | 2022-07-22 | 2022-12-09 | 中国人民解放军战略支援部队航天工程大学 | Multi-source information fusion uncertainty judgment method |
| CN115457351B (en) * | 2022-07-22 | 2023-10-20 | 中国人民解放军战略支援部队航天工程大学 | Multi-source information fusion uncertainty judging method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107992889A (en) | A kind of iron spectrogram based on D-S evidence theory is as Multi-information acquisition method | |
| EP3452959B1 (en) | Model construction in a neural network for object detection | |
| Hyeon et al. | Diagnosing cervical cell images using pre-trained convolutional neural network as feature extractor | |
| CN105005765A (en) | Facial expression identification method based on Gabor wavelet and gray-level co-occurrence matrix | |
| WO2013037119A1 (en) | Device and method for erythrocyte morphology analysis | |
| CN110502972A (en) | A kind of pavement crack segmentation and recognition methods based on deep learning | |
| CN111008647B (en) | Sample extraction and image classification method based on void convolution and residual linkage | |
| CN103927758B (en) | Saliency detection method based on contrast ratio and minimum convex hull of angular point | |
| CN103106265A (en) | Method and system of classifying similar images | |
| CN104182985A (en) | Remote sensing image change detection method | |
| CN106548131A (en) | A kind of workmen's safety helmet real-time detection method based on pedestrian detection | |
| CN105243387A (en) | Open-pit mine typical ground object classification method based on UAV image | |
| CN103413124A (en) | Method for detecting round traffic sign | |
| CN103500342A (en) | Human behavior recognition method based on accelerometer | |
| CN107818321A (en) | A kind of watermark date recognition method for vehicle annual test | |
| CN107657220A (en) | A kind of leukorrhea mould automatic testing method based on HOG features and SVM | |
| Li et al. | GIS-based detection of grain boundaries | |
| CN104573701B (en) | A kind of automatic testing method of Tassel of Corn | |
| CN112581450A (en) | Pollen detection method based on expansion convolution pyramid and multi-scale pyramid | |
| CN112784767A (en) | Cell example segmentation algorithm based on leukocyte microscopic image | |
| CN115620178A (en) | Real-time detection method for abnormal and dangerous behaviors of power grid of unmanned aerial vehicle | |
| CN106548195A (en) | A kind of object detection method based on modified model HOG ULBP feature operators | |
| CN105469099A (en) | Sparse-representation-classification-based pavement crack detection and identification method | |
| Sulaiman et al. | Semi-automated pseudo colour features extraction technique for cervical cancer's pap smear images | |
| Nithya et al. | Detection of Anaemia using Image Processing Techniques from microscopy blood smear images |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180504 |