CN108846307A - A kind of microseism based on waveform image and explosion events recognition methods - Google Patents
A kind of microseism based on waveform image and explosion events recognition methods Download PDFInfo
- Publication number
- CN108846307A CN108846307A CN201810326479.6A CN201810326479A CN108846307A CN 108846307 A CN108846307 A CN 108846307A CN 201810326479 A CN201810326479 A CN 201810326479A CN 108846307 A CN108846307 A CN 108846307A
- Authority
- CN
- China
- Prior art keywords
- waveform
- image
- event
- waveform image
- microseism
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/02—Preprocessing
-
- 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
- G06F18/211—Selection of the most significant subset of features
-
- 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
- G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
- G06F2218/10—Feature extraction by analysing the shape of a waveform, e.g. extracting parameters relating to peaks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Signal Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of microseisms based on waveform image and explosion events recognition methods, include the following steps:Step 1:Establish the waveform image database of microseism and explosion events:Step 2:All kinds of event characteristics of image obtain:Original waveform characteristics of image is extracted to M group microseismic event and N group explosion events using Principal Component Analysis (PCA), then it reduces data dimension and eliminates the correlation between different characteristic, simultaneous quantitative retains the most useful characteristic information for including in original waveform characteristics of image;Step 3:Disaggregated model is established:Using the waveform image feature after the acquired dimensionality reduction of machine learning algorithm LIBSVM training, and establish the disaggregated model of microseism and explosion events;Step 4:Identification events are treated to be identified:The waveform image feature for inputting event to be identified, treats identification events with explosion events disaggregated model according to established microseism and classifies.The present invention have the characteristics that it is widely applicable, accurate quickly, objectivity it is strong.
Description
Technical field
The invention belongs to micro seismic monitoring field, in particular to a kind of microseism based on waveform image and explosion events identification side
Method.
Background technique
Micro seismic monitoring is at home and abroad widely applied as a kind of effective ground pressure monitoring means, and microseism and explosion
The Division identification of event is the critical issue in data handling procedure, therefore is had to the research of microseism and explosion events recognition methods
It is significant.However, for the Microseismic monitoring system of normal work, the microseism and explosion events quantity that monitor daily
It is extremely more, there is similitude between different event, by the differentiating method of traditional artificial identification vulnerable to operator's professional knowledge and
The influence of experience not only identifies limited amount, but also is likely to cause recognition result disunity to have serious consequences, and reduces
The effect that micro seismic monitoring is analyzed in real time.
Knowing method for distinguishing with explosion events for microseism at present can be divided mainly into:Waveform frequency spectrum analytic approach, multi-parameter statistics
Method and machine learning method, wherein multi-parameter statistic law includes Gauss maximum-likelihood method, fisher classification, the Logistic Return Law
Deng.Machine learning neural network, support vector machines method, random forest method etc..These methods generally include feature extraction and spy
Sign two processes of identification, common feature extraction is in two aspects:First is that focal shock parameter, second is that waveform parameter.Common focus
Number of parameters is more, and can derive more focal shock parameters by the fitting or conversion of different functions, is selected by statistical analysis
It is extremely time-consuming and laborious to select the focal shock parameter with preferable recognition capability.Meanwhile the different focal shock parameters with preferable recognition capability
Between combination also need test of many times and verifying, and be most likely subject to the influence of experience and subjective judgement.For waveform parameter
For, leading portion waveform is mostly just considered, and have ignored Full wave shape and the good effect played is identified to microseism and explosion events.
In addition, the importance of usually selected waveform parameter does not arrive during establishing disaggregated model, the phase between different parameters
Closing property is not eliminated, and classifying quality is affected.
It can be seen that existing microseismic event and explosion events recognition methods need to study a kind of applicability there are biggish limitation
By force, the high automatic identifying method of accuracy rate.
Summary of the invention
The technical problem to be solved by the present invention is in view of the deficiencies of the prior art, provide a kind of based on waveform image
Microseism and explosion events recognition methods, the microseism and explosion events recognition methods are widely applicable, accurate quick.
A kind of microseism based on waveform image and explosion events recognition methods, include the following steps:
Step 1:Construct the waveform image sample database of microseism and explosion events;
Using known microseismic event waveform image corresponding with the generation of the all-wave graphic data of explosion events, M is randomly selected
The waveform image of group microseismic event and N group explosion events counts waveform image sample number as waveform image sample database
The number and accounting occurred in entire waveform sample data library according to the waveform time length of waveform image each in library, wherein M
It is integer, M=N, and M, N >=100 with N;
The ratio of the number that occurs in entire waveform sample data library using waveform time length and (M+N) are as correspondence wave
The accounting of shape time span;
Step 2:When constructing the optimum waveform of microseism and explosion events in waveform image sample database according to the following conditions
Between length model, tsThe optimum waveform time span of expression event;
η(t1)+η(t2)+η(t3)+…+η(tn) >=80%
Wherein, tsThe optimum waveform time span of expression event;To waveform time length in wave pattern database
After the descending arrangement of the number of appearance, tiIt indicates in waveform image sample database, the waveform time for coming i-th bit is long
Degree;η(t1)、η(t2)、η(t3) and η (tn) refer respectively in waveform image sample database, the 1st, 2,3 and n waveforms
The accounting of time span;N is indicated so that formula η (t1)+η(t2)+η(t3)+…+η(tnThe smallest positive integral that) >=80% is set up;
Step 3:Based on optimum waveform time span, at the carry out shaping of waveform image in waveform image sample database
Reason;
Step 4:All microseismic event sample waveform images and explosion events sample after Shape correction are extracted respectively
The characteristics of image of waveform image;
The extraction of characteristics of image is to extract to the gray value of each pixel in every width waveform image, obtains one
Then a two-dimensional matrix converts row vector for two-dimensional matrix, then the characteristics of image of all samples in every a kind of event is constituted
One two-dimensional matrix;
Step 5:Construct microseism and explosion events disaggregated model;
Using the characteristics of image of all event samples as input data, corresponding event category label as output data,
LIBSVM model based on machine learning is trained, microseism and explosion events disaggregated model are obtained;
Step 6:The waveform image feature for extracting event to be identified utilizes the microseism built and explosion events classification mould
Type carries out event type identification;
Wherein, the extraction process of the waveform image feature of event to be identified is first to extract waveform diagram from Full wave shape image
Then picture carries out the image characteristics extraction of pixel gray value to waveform image.
Further, when the carry out Shape correction to waveform image in waveform image sample database refers to waveform
Between length t be less than optimal time length tsEvent, take (ts- t) portion waveshape speed be 0, i.e. y=0;It is long to waveform time
It spends t and is greater than optimal time length tsEvent, take [0, ts] portion waveshape;
Take optimum waveform time span tsIt is that the most microseisms or explosion established are guaranteed according to statistical analysis first
The waveform image of event all remains original waveform to the greatest extent, to reflect the feature of respective event;Unified waveform image rule
Lattice avoid established waveform image by cross directional stretch or compression, original wave character are caused to change, and to final
Recognition result impacts.
Further, the waveform image feature of sample waveform characteristics of image and event to be identified is mentioned using principal component analysis method
The process is taken to be:
A two-dimensional matrix is converted by the gray value of pixel each in each waveform image, and converts two-dimensional matrix to
One row vector x;
Wherein, xpqIt is q-th of feature of p-th of waveform image, p is the quantity of event in sample database, and q is waveform diagram
Characteristic Number as in.
Further, dimension-reduction treatment is carried out using principal component analysis method to the waveform image feature of extraction, detailed process is as follows:
Step A:It averages to the pixel value of the same pixel of different wave image, seeks XpqIn each column be averaged
Value, obtains a line vector xave, acquire matrix of differences Xd;
Which removes the similitudes between different wave image to remain otherness simultaneously;
Step B:Calculating difference matrix XdCovariance matrix Cqq, then calculate covariance matrix CqqEigenvalue λ, corresponding
Characteristic value weight w (λ) and corresponding feature vector e;
Step C:Characteristic value is ranked up by characteristic value size descending, obtains λ1≥λ2≥…≥λq, utilize the tribute of setting
The rate σ of offering determines Data Dimensionality Reduction degree:
w(λ1)+w(λ2)+…+w(λk)≥σ
Wherein, k is the smallest positive integral for meeting above formula, k<Q, 0<σ<1;
Step D:The k column before choosing in the eigenvectors matrix that the feature vector e that step B is obtained is constituted, obtain dimensionality reduction square
Battle array XDR, and utilize original waveform image characteristic matrix XpqWith dimensionality reduction matrix XDRIt is multiplied, the waveform image feature square after obtaining dimensionality reduction
Battle array XR:
XR=Xpq·XDR。
Covariance matrix C is acquiredqqEigenvalue λ and corresponding feature vector e after, can be obtained a matrix
K is acquired according to contribution rate, k column, obtain dimensionality reduction matrix before selecting in matrix as above.
For sample waveform characteristics of image after dimensionality reduction for constructing disaggregated model, the event waveforms image to be identified after dimensionality reduction is special
Sign input disaggregated model;
Further, selection linear kernel function instructs the LIBSVM model based on machine learning described in step 5
Practice.
Further, the optimum waveform time span ts=1.8s.
Further, contribution rate σ=0.95.
Beneficial effect
The present invention provides a kind of microseisms based on waveform image and explosion events recognition methods, are based on different types of thing
Part has different waveform image features, so that extract waveform image is indicated with matrix, and right using Principal Component Analysis (PCA)
Original waveform characteristics of image carries out dimensionality reduction to improve classification effectiveness, and quantitatively include in reservation original waveform characteristics of image most has
Characteristic information.In conjunction with the waveform image feature after machine learning algorithm LIBSVM and dimensionality reduction, microseism and explosion events are established
Disaggregated model.By inputting the waveform image feature of event to be identified, the classification of Accurate Prediction event.Realize micro seismic monitoring
Real-time analysis, alleviate the data processing pressure of staff, weaken influence of the human factor to monitoring result, improve
Accuracy.The method has the characteristics that using extensive, accurate quick.
Detailed description of the invention
Fig. 1 is the flow chart of the method for the invention;
Fig. 2 is waveform image feature schematic diagram.
Specific embodiment
Below in conjunction with drawings and examples, the present invention is described further.
As shown in Figure 1, a kind of microseism based on waveform image and explosion events recognition methods, include the following steps:
Step 1:Construct the waveform image sample database of microseism and explosion events;
Using known microseismic event waveform image corresponding with the generation of the all-wave graphic data of explosion events, M is randomly selected
The waveform image of group microseismic event and N group explosion events counts waveform image sample number as waveform image sample database
The number and accounting occurred in entire waveform sample data library according to the waveform time length of waveform image each in library, wherein M
It is integer, M=N, and M, N >=100 with N;
The ratio of the number that occurs in entire waveform sample data library using waveform time length and (M+N) are as correspondence wave
The accounting of shape time span;
Step 2:When constructing the optimum waveform of microseism and explosion events in waveform image sample database according to the following conditions
Between length model, tsThe optimum waveform time span of expression event;
η(t1)+η(t2)+η(t3)+…+η(tn) >=80%
Wherein, tsThe optimum waveform time span of expression event;To waveform time length in wave pattern database
After the descending arrangement of the number of appearance, tiIt indicates in waveform image sample database, the waveform time for coming i-th bit is long
Degree;η(t1)、η(t2)、η(t3) and η (tn) refer respectively in waveform image sample database, the 1st, 2,3 and n waveforms
The accounting of time span;N is indicated so that formula η (t1)+η(t2)+η(t3)++η(tnThe smallest positive integral that) >=80% is set up;
In this example, optimum waveform time span ts=1.8s.
Step 3:Based on optimum waveform time span, at the carry out shaping of waveform image in waveform image sample database
Reason;
The carry out Shape correction of waveform image in waveform image sample database is referred to, waveform time length t is less than most
Excellent time span tsEvent, take (ts- t) portion waveshape speed be 0, i.e. y=0;When being greater than optimal to waveform time length t
Between length tsEvent, take [0, ts] portion waveshape;
Take optimum waveform time span tsIt is that the most microseisms or explosion established are guaranteed according to statistical analysis first
The waveform image of event all remains original waveform to the greatest extent, to reflect the feature of respective event;Unified waveform image rule
Lattice avoid established waveform image by cross directional stretch or compression, original wave character are caused to change, and to final
Recognition result impacts.
Step 4:All microseismic event sample waveform images and explosion events sample after Shape correction are extracted respectively
The characteristics of image of waveform image;
As shown in Fig. 2, the extraction of characteristics of image, is mentioned to the gray value of each pixel in every width waveform image
It takes, obtains a two-dimensional matrix, then convert row vector for two-dimensional matrix, then by the image of all samples in every a kind of event
Feature constitutes a two-dimensional matrix;
The enlarged drawing of Blocked portion waveform is right part of flg, for show pixel distribution and corresponding gray value,
The picture for scheming the 0-255 in following face is gray value profiles, can be corresponded the gray value of each pixel by this,
To extract feature.
Waveform image characteristic extraction procedure to sample waveform characteristics of image and event to be identified is:
A two-dimensional matrix is converted by the gray value of pixel each in each waveform image, and converts two-dimensional matrix to
One row vector x;
Wherein, xpqIt is q-th of feature of p-th of waveform image, p is the quantity of event in sample database, and q is waveform diagram
Characteristic Number as in.
Dimension-reduction treatment is carried out using principal component analysis method to the waveform image feature of extraction, detailed process is as follows:
Step A:It averages to the pixel value of the same pixel of different wave image, seeks XpqIn each column be averaged
Value, obtains a line vector xave, acquire matrix of differences Xd;
Which removes the similitudes between different wave image to remain otherness simultaneously;
Step B:Calculating difference matrix XdCovariance matrix Cqq, then calculate covariance matrix CqqEigenvalue λ, corresponding
Characteristic value weight w (λ) and corresponding feature vector e;
Step C:Characteristic value is ranked up by characteristic value size descending, obtains λ1≥λ2≥…≥λq, utilize the tribute of setting
The rate σ of offering determines Data Dimensionality Reduction degree:
w(λ1)+w(λ2)+…+w(λk)≥σ
Wherein, k is the smallest positive integral for meeting above formula, k<Q, 0<σ<1;In this example, contribution rate σ=0.95.
Step D:The k column before choosing in the eigenvectors matrix that the feature vector e that step B is obtained is constituted, obtain dimensionality reduction square
Battle array XDR, and utilize original waveform image characteristic matrix XpqWith dimensionality reduction matrix XDRIt is multiplied, the waveform image feature square after obtaining dimensionality reduction
Battle array XR:
XR=Xpq·XDR。
Covariance matrix C is acquiredqqEigenvalue λ and corresponding feature vector e after, can be obtained a matrix
K is acquired according to contribution rate, k column, obtain dimensionality reduction matrix before selecting in matrix as above.
For sample waveform characteristics of image after dimensionality reduction for constructing disaggregated model, the event waveforms image to be identified after dimensionality reduction is special
Sign input disaggregated model;
Step 5:Construct microseism and explosion events disaggregated model;
Using the characteristics of image of all event samples as input data, corresponding event category label as output data,
LIBSVM model based on machine learning is trained, microseism and explosion events disaggregated model are obtained;
Selection linear kernel function is trained the LIBSVM model based on machine learning described in step 5
Step 6:The waveform image feature for extracting event to be identified utilizes the microseism built and explosion events classification mould
Type carries out event type identification;
Wherein, the extraction process of the waveform image feature of event to be identified is first to extract waveform diagram from Full wave shape image
Then picture carries out the image characteristics extraction of pixel gray value to waveform image.
Embodiment 1:
(1) fixed 1000 groups of microseismic events and 1000 groups of explosion events are obtained by Microseismic monitoring system, taken optimal
Waveform time length ts=1.8s establishes microseism and explosion events waveform image database.
(2) contribution rate σ=0.95, former characteristic dimension=2000, dimensionality reduction are taken described in middle step 2 according to claim 1
Characteristic dimension=1157 afterwards, see Table 1 for details.
Characteristic after dimensionality reduction eliminates different wave image similarity, remains otherness, eliminates former different special
Correlation between sign remains the characteristic information for including in preceding 95% original waveform characteristics of image, improves classification effectiveness.
1 principal component analysis result of table
(3) characteristic and machine learning algorithm LIBSVM after combining dimensionality reduction, two points for establishing microseism and explosion events
Class model, Parameters in Mathematical Model are as shown in table 2.
The major parameter of table 2 microseism and explosion events disaggregated model
Wherein:Niter:The number of iterations
obj:The optimal value of Quadratic Programming Solution
rho:The bias term of discriminant function
NSV:Standard supporting vector number
NBSV:Borderline supporting vector number
Total NSV:Supporting vector total number
(4) the waveform image feature of fixed other 1000 groups of microseismic events and other 1000 groups of explosion events is inputted,
Examine recognition accuracy:
The microseismic event quantity correctly identified:950 accuracys rate:95.0%
The explosion events quantity correctly identified:922 accuracys rate:92.2%
Average recognition accuracy:(95.0%+92.2%)/2=93.6%
The Average Accuracy of microseism and explosion events identification meets practical implementation up to 93.6%, up to 95.0%
The required accuracy.
The above description is only an embodiment of the present invention, is not intended to limit the invention, all in spirit of that invention and original
It within then, changes, equivalent replacement, improvement etc., should all be included in the protection scope of the present invention.
Claims (7)
1. a kind of microseism based on waveform image and explosion events recognition methods, which is characterized in that include the following steps:
Step 1:Construct the waveform image sample database of microseism and explosion events;
Using known microseismic event waveform image corresponding with the generation of the all-wave graphic data of explosion events, it is micro- to randomly select M group
The waveform image of shake event and N group explosion events counts waveform image sample database as waveform image sample database
In each waveform image the number that occurs in entire waveform sample data library of waveform time length and accounting, wherein M and N
It is integer, M=N, and M, N >=100;
The ratio of the number and (M+N) that occur in entire waveform sample data library using waveform time length is as when corresponding to waveform
Between length accounting;
Step 2:It is long according to the optimum waveform time of microseism and explosion events in the following conditions building waveform image sample database
Spend model, tsThe optimum waveform time span of expression event;
Wherein, tsThe optimum waveform time span of expression event;Occur in wave pattern database to waveform time length
After the descending arrangement of number, tiIt indicates in waveform image sample database, comes the waveform time length of i-th bit;η(t1)、
η(t2)、η(t3) and η (tn) refer respectively in waveform image sample database, the 1st, 2,3 and n waveform time length
Accounting;N is indicated so that formula η (t1)+η(t2)+η(t3)+…+η(tnThe smallest positive integral that) >=80% is set up;
Step 3:Based on optimum waveform time span, to the carry out Shape correction of waveform image in waveform image sample database;
Step 4:All microseismic event sample waveform images and explosion events sample waveform after Shape correction are extracted respectively
The characteristics of image of image;
Step 5:Construct microseism and explosion events disaggregated model;
Using the characteristics of image of all event samples as input data, corresponding event category label is as output data, to base
It is trained in the LIBSVM model of machine learning, obtains microseism and explosion events disaggregated model;
Step 6:The waveform image feature for extracting event to be identified, using the microseism that builds and explosion events disaggregated model, into
Act part type identification;
Wherein, the extraction process of the waveform image feature of event to be identified is that waveform image is first extracted from Full wave shape image, so
The image characteristics extraction of pixel gray value is carried out to waveform image afterwards.
2. the method according to claim 1, wherein described to waveform image in waveform image sample database
Shape correction is carried out to refer to waveform time length t less than optimal time length tsEvent, take (ts- t) portion waveshape speed
It is 0, i.e. y=0;Optimal time length t is greater than to waveform time length tsEvent, take [0, ts] portion waveshape.
3. the method according to claim 1, wherein using principal component analysis method to sample waveform characteristics of image and to
The waveform image characteristic extraction procedure of identification events is:
A two-dimensional matrix is converted by the gray value of pixel each in each waveform image, and converts one for two-dimensional matrix
Row vector x;
Wherein, xpqIt is q-th of feature of p-th of waveform image, p is the quantity of event in sample database, and q is in waveform image
Characteristic Number.
4. according to the method described in claim 3, it is characterized in that, to the waveform image feature of extraction using principal component analyse method into
Row dimension-reduction treatment, detailed process is as follows:
Step A:It averages to the pixel value of the same pixel of different wave image, seeks XpqIn each column average value, obtain
A line vector xave, acquire matrix of differences Xd;
Step B:Calculating difference matrix XdCovariance matrix Cqq, then calculate covariance matrix CqqEigenvalue λ, corresponding feature
It is worth weight w (λ) and corresponding feature vector e;
Step C:Characteristic value is ranked up by characteristic value size descending, obtains λ1≥λ2≥…≥λq, utilize the contribution rate of setting
σ determines Data Dimensionality Reduction degree:
w(λ1)+w(λ2)+…+w(λk)≥σ
Wherein, k is the smallest positive integral for meeting above formula, k<Q, 0<σ<1;
Step D:The k column before choosing in the eigenvectors matrix that the feature vector e that step B is obtained is constituted, obtain dimensionality reduction matrix XDR,
And utilize original waveform image characteristic matrix XpqWith dimensionality reduction matrix XDRIt is multiplied, the waveform image eigenmatrix X after obtaining dimensionality reductionR:
XR=Xpq·XDR。
5. the method according to claim 1, wherein selection linear kernel function to described in step 5 based on machine
The LIBSVM model of device study is trained.
6. the method according to claim 1, wherein the optimum waveform time span ts=1.8s.
7. according to the described in any item methods of claim 4-6, which is characterized in that contribution rate σ=0.95.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810326479.6A CN108846307B (en) | 2018-04-12 | 2018-04-12 | Microseism and blasting event identification method based on waveform image |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810326479.6A CN108846307B (en) | 2018-04-12 | 2018-04-12 | Microseism and blasting event identification method based on waveform image |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108846307A true CN108846307A (en) | 2018-11-20 |
CN108846307B CN108846307B (en) | 2021-12-28 |
Family
ID=64211962
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810326479.6A Active CN108846307B (en) | 2018-04-12 | 2018-04-12 | Microseism and blasting event identification method based on waveform image |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108846307B (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110058294A (en) * | 2019-05-10 | 2019-07-26 | 东北大学 | A kind of tunnel micro seismic monitoring rock rupture event automatic identifying method |
CN110414723A (en) * | 2019-07-09 | 2019-11-05 | 中国石油大学(北京) | The method, apparatus and system of fractured hydrocarbon reservoir history matching based on microseismic event |
CN112257560A (en) * | 2020-10-20 | 2021-01-22 | 华北电力大学 | Microseismic event detection method and system by utilizing cumulative similarity |
CN112487952A (en) * | 2020-11-27 | 2021-03-12 | 东北大学 | Mine microseismic signal automatic identification method based on deep learning |
CN115330779A (en) * | 2022-10-13 | 2022-11-11 | 四川迪晟新达类脑智能技术有限公司 | Blasting time confirmation method and system based on fire light and smoke dust in civil blasting |
Citations (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101118592A (en) * | 2007-08-22 | 2008-02-06 | 大连理工大学 | Printers evidence obtaining method based on character printing feature |
US20110082647A1 (en) * | 2009-10-05 | 2011-04-07 | Pascal Edme | Combining seismic data from sensors to attenuate noise |
CN103109207A (en) * | 2010-08-30 | 2013-05-15 | 麦克罗地震探测公司 | Method for detection of subsurface seismic events in vertically transversely isotropic media |
CN103410569A (en) * | 2013-08-27 | 2013-11-27 | 武汉钢铁(集团)公司 | Metal mine downhole microseismic monitoring system |
CN103605151A (en) * | 2013-11-20 | 2014-02-26 | 中北大学 | Distributed group wave shallow-layer slight shock positioning method based on phase measuring |
CN104297788A (en) * | 2014-10-20 | 2015-01-21 | 中南大学 | Mine microseism and blasting signal identification method based on waveform oscillation starting trend line slope |
CN104834004A (en) * | 2015-04-13 | 2015-08-12 | 中南大学 | Mine slight shock and blasting signal identification method based on waveform slope before and after peak value |
CN105740840A (en) * | 2016-02-29 | 2016-07-06 | 中南大学 | Nonlinear identification method for rock fracture signal and blasting vibration signal |
CN106202919A (en) * | 2016-07-08 | 2016-12-07 | 中南大学 | A kind of microseism based on focal shock parameter and explosion events recognition methods |
US20170168177A1 (en) * | 2015-12-11 | 2017-06-15 | Ion Geophysical Corporation | System and method for reconstructed wavefield inversion |
CN106874833A (en) * | 2016-12-26 | 2017-06-20 | 中国船舶重工集团公司第七0研究所 | A kind of mode identification method of vibration event |
CN107292246A (en) * | 2017-06-05 | 2017-10-24 | 河海大学 | Infrared human body target identification method based on HOG PCA and transfer learning |
CN107505652A (en) * | 2017-07-26 | 2017-12-22 | 山东科技大学 | A kind of mine microquake signal discrimination method based on energy-distributing feature |
CN107527045A (en) * | 2017-09-19 | 2017-12-29 | 桂林安维科技有限公司 | A kind of human body behavior event real-time analysis method towards multi-channel video |
CN107582077A (en) * | 2017-08-17 | 2018-01-16 | 南京信息工程大学 | A kind of human body state of mind analysis method that behavior is touched based on mobile phone |
CN107844865A (en) * | 2017-11-20 | 2018-03-27 | 天津科技大学 | Feature based parameter chooses the stock index prediction method with LSTM models |
-
2018
- 2018-04-12 CN CN201810326479.6A patent/CN108846307B/en active Active
Patent Citations (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101118592A (en) * | 2007-08-22 | 2008-02-06 | 大连理工大学 | Printers evidence obtaining method based on character printing feature |
US20110082647A1 (en) * | 2009-10-05 | 2011-04-07 | Pascal Edme | Combining seismic data from sensors to attenuate noise |
CN103109207A (en) * | 2010-08-30 | 2013-05-15 | 麦克罗地震探测公司 | Method for detection of subsurface seismic events in vertically transversely isotropic media |
CN103410569A (en) * | 2013-08-27 | 2013-11-27 | 武汉钢铁(集团)公司 | Metal mine downhole microseismic monitoring system |
CN103605151A (en) * | 2013-11-20 | 2014-02-26 | 中北大学 | Distributed group wave shallow-layer slight shock positioning method based on phase measuring |
CN104297788A (en) * | 2014-10-20 | 2015-01-21 | 中南大学 | Mine microseism and blasting signal identification method based on waveform oscillation starting trend line slope |
CN104834004A (en) * | 2015-04-13 | 2015-08-12 | 中南大学 | Mine slight shock and blasting signal identification method based on waveform slope before and after peak value |
US20170168177A1 (en) * | 2015-12-11 | 2017-06-15 | Ion Geophysical Corporation | System and method for reconstructed wavefield inversion |
CN105740840A (en) * | 2016-02-29 | 2016-07-06 | 中南大学 | Nonlinear identification method for rock fracture signal and blasting vibration signal |
CN106202919A (en) * | 2016-07-08 | 2016-12-07 | 中南大学 | A kind of microseism based on focal shock parameter and explosion events recognition methods |
CN106874833A (en) * | 2016-12-26 | 2017-06-20 | 中国船舶重工集团公司第七0研究所 | A kind of mode identification method of vibration event |
CN107292246A (en) * | 2017-06-05 | 2017-10-24 | 河海大学 | Infrared human body target identification method based on HOG PCA and transfer learning |
CN107505652A (en) * | 2017-07-26 | 2017-12-22 | 山东科技大学 | A kind of mine microquake signal discrimination method based on energy-distributing feature |
CN107582077A (en) * | 2017-08-17 | 2018-01-16 | 南京信息工程大学 | A kind of human body state of mind analysis method that behavior is touched based on mobile phone |
CN107527045A (en) * | 2017-09-19 | 2017-12-29 | 桂林安维科技有限公司 | A kind of human body behavior event real-time analysis method towards multi-channel video |
CN107844865A (en) * | 2017-11-20 | 2018-03-27 | 天津科技大学 | Feature based parameter chooses the stock index prediction method with LSTM models |
Non-Patent Citations (2)
Title |
---|
SAEED GHORBANI等: "A Modern Method to Improve of Detecting and Categorizing Mechanism for Micro Seismic Events Data Using Boost Learning System", 《CIVIL ENGINEERING JOURNAL》 * |
董陇军等: "微震与***事件统计识别方法及工程应用", 《岩石力学与工程学报》 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110058294A (en) * | 2019-05-10 | 2019-07-26 | 东北大学 | A kind of tunnel micro seismic monitoring rock rupture event automatic identifying method |
CN110414723A (en) * | 2019-07-09 | 2019-11-05 | 中国石油大学(北京) | The method, apparatus and system of fractured hydrocarbon reservoir history matching based on microseismic event |
CN112257560A (en) * | 2020-10-20 | 2021-01-22 | 华北电力大学 | Microseismic event detection method and system by utilizing cumulative similarity |
CN112487952A (en) * | 2020-11-27 | 2021-03-12 | 东北大学 | Mine microseismic signal automatic identification method based on deep learning |
CN115330779A (en) * | 2022-10-13 | 2022-11-11 | 四川迪晟新达类脑智能技术有限公司 | Blasting time confirmation method and system based on fire light and smoke dust in civil blasting |
CN115330779B (en) * | 2022-10-13 | 2022-12-20 | 四川迪晟新达类脑智能技术有限公司 | Blasting time confirmation method and system based on fire light and smoke dust in civil blasting |
Also Published As
Publication number | Publication date |
---|---|
CN108846307B (en) | 2021-12-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108846307A (en) | A kind of microseism based on waveform image and explosion events recognition methods | |
CN108537743B (en) | Face image enhancement method based on generation countermeasure network | |
Xu et al. | Maize diseases identification method based on multi-scale convolutional global pooling neural network | |
Lee et al. | YASS: yet another spike sorter | |
CN108388896A (en) | A kind of licence plate recognition method based on dynamic time sequence convolutional neural networks | |
CN104778481A (en) | Method and device for creating sample library for large-scale face mode analysis | |
CN106778714B (en) | LDA face identification method based on nonlinear characteristic and model combination | |
CN101996328B (en) | Wood identification method | |
CN106845497B (en) | Corn early-stage image drought identification method based on multi-feature fusion | |
JP2002520699A (en) | Non-literal pattern recognition method and system for hyperspectral image utilization | |
CN106295694A (en) | Face recognition method for iterative re-constrained group sparse representation classification | |
CN112241762A (en) | Fine-grained identification method for pest and disease damage image classification | |
CN109214298A (en) | A kind of Asia women face value Rating Model method based on depth convolutional network | |
CN103177266A (en) | Intelligent stock pest identification system | |
CN107563427A (en) | The method and corresponding use that copyright for oil painting is identified | |
CN110533100A (en) | A method of CME detection and tracking is carried out based on machine learning | |
CN110502989A (en) | A kind of small sample EO-1 hyperion face identification method and system | |
CN102314591B (en) | Method and equipment for detecting static foreground object | |
Fang et al. | Multiscale CNNs ensemble based self-learning for hyperspectral image classification | |
CN106326914B (en) | A kind of more classification methods of pearl based on SVM | |
Ullah et al. | Automatic diseases detection and classification in maize crop using convolution neural network | |
CN113077452A (en) | Apple tree pest and disease detection method based on DNN network and spot detection algorithm | |
CN111310711B (en) | Face image recognition method and system based on two-dimensional singular spectrum analysis and EMD fusion | |
CN110717691B (en) | Assembly method and system for evaluation model based on components | |
CN105139019A (en) | Iris image screening method and device |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |