CN106951924A - Acceleration Algorithm in Seismic Coherence Cube image slices automatic identifying method and system based on AdaBoost algorithms - Google Patents
Acceleration Algorithm in Seismic Coherence Cube image slices automatic identifying method and system based on AdaBoost algorithms Download PDFInfo
- Publication number
- CN106951924A CN106951924A CN201710186136.XA CN201710186136A CN106951924A CN 106951924 A CN106951924 A CN 106951924A CN 201710186136 A CN201710186136 A CN 201710186136A CN 106951924 A CN106951924 A CN 106951924A
- Authority
- CN
- China
- Prior art keywords
- image block
- training sample
- image
- acceleration algorithm
- sorter network
- 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
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/285—Selection of pattern recognition techniques, e.g. of classifiers in a multi-classifier system
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention provides a kind of Acceleration Algorithm in Seismic Coherence Cube image slices automatic identifying method and system based on AdaBoost algorithms, method is：The image block in Acceleration Algorithm in Seismic Coherence Cube faultage image is obtained, image block includes fault body image block, tomography edge image block and tomography background image block；By the sorter network of training in advance, image block is classified, classification results are obtained, sorter network is made up of multiple graders, and sorter network is obtained by AdaBoost Algorithm Learnings；According to classification results, the automatic identification to Acceleration Algorithm in Seismic Coherence Cube faultage image is realized.Sorter network is trained present invention employs the method for integrated study, image block is classified with sorter network, the automatic identification that can make tomography and the result extracted are more accurate.
Description
Technical field
The present invention relates to a kind of Acceleration Algorithm in Seismic Coherence Cube fault recognizing field, more particularly to the seismic facies based on AdaBoost algorithms
Stem body image slices automatic identifying method and system.
Background technology
Seismic coherence interpretation is the main technology for realizing geological fault position and attributive analysis.However, due to earthquake number
According to inevitably existing during field acquisition due to equipment, environment and the external factor such as artificial, and these it is external because
Element influences on complex jamming and noise produced by the effect of different geologic structures etc., and these data for causing of influence degrade it is past
Past is nonlinear.Therefore, the complexity of stratum background and tomography target data often to be randomly distributed in earthquake fault image
Data.It is traditional based on methods such as Image Edge-Detection/contours extract, image segmentation/Objective extractions in face of this complex data
Accurately identifying and positioning for tomography difficult to realize, generally requires more manual editing work.In addition, in Acceleration Algorithm in Seismic Coherence Cube image
Tomography and background classes ratio data are uneven, and the openness picture of background data is to preferable, and tomography is relatively complicated, is automatic tomography
Identification brings certain difficulty.
In view of artificial intelligence approaches such as machine learning in the related every subjects of Signal and Information Processing and field
In be obtained for preferable application.Particularly, integrated learning approach the prior distribution of training data is not carried out it is any it is assumed that
There is very strong robustness to data distribution situation.Such method, can by directly working in the probability space of input data
Effectively solve the uneven classification task of class imbalance, attribute.Obviously, it is necessary to carry out integrated study side from the angle of classification
Application technology of the method in Acceleration Algorithm in Seismic Coherence Cube fault recognizing.
Therefore, defect of the prior art is, earthquake fault method in existing Acceleration Algorithm in Seismic Coherence Cube image, due to tomography point
Cloth is irregular, the complex-shaped changeable, tomography of tomography and background data ratio imbalance etc. cause automatic identification and the extraction of tomography
Result it is inaccurate.
The content of the invention
For above-mentioned technical problem, the present invention provides a kind of Acceleration Algorithm in Seismic Coherence Cube image slices based on AdaBoost algorithms certainly
Dynamic recognition methods and system, employ the method training sorter network of integrated study, image block are divided with sorter network
Class, the automatic identification that can make tomography and the result extracted are more accurate.
In order to solve the above technical problems, the technical scheme that the present invention is provided is：
In a first aspect, the present invention provides a kind of Acceleration Algorithm in Seismic Coherence Cube image slices automatic identification side based on AdaBoost algorithms
Method, including：
Step S1, obtains the image block in Acceleration Algorithm in Seismic Coherence Cube faultage image, and described image block includes fault body image block, broken
Layer edge image block and tomography background image block；
Step S2, by the sorter network of training in advance, classifies to image block, obtains classification results, the classification
Network is made up of multiple graders, and the sorter network is obtained by AdaBoost Algorithm Learnings；
Step S3, according to the classification results, realizes the automatic identification to the Acceleration Algorithm in Seismic Coherence Cube faultage image.
The Acceleration Algorithm in Seismic Coherence Cube image slices automatic identifying method based on AdaBoost algorithms of the present invention, its technical scheme
It is：The image block in Acceleration Algorithm in Seismic Coherence Cube faultage image is obtained, described image block includes fault body image block, tomography edge image block
With tomography background image block；By the sorter network of training in advance, image block is classified, classification results, described point are obtained
Class network is made up of multiple graders, and the sorter network is obtained by AdaBoost Algorithm Learnings；According to the classification results,
Realize the automatic identification to the Acceleration Algorithm in Seismic Coherence Cube faultage image.
The Acceleration Algorithm in Seismic Coherence Cube image slices automatic identifying method based on AdaBoost algorithms of the present invention, employs integrated
The method training sorter network of habit, classifies with sorter network to image block, can make automatic identification and the extraction of tomography
As a result it is more accurate.
Further, the training process of the sorter network is：
The first training sample of N number of off-line training sample formation is chosen, the study to first training sample obtains first
Individual Weak Classifier, the off-line training sample is the image block in Acceleration Algorithm in Seismic Coherence Cube faultage image；
The sample of misclassification in first Weak Classifier and other off-line training samples are constituted into second training sample,
Second Weak Classifier is obtained by the study to second training sample, second training sample is by N number of offline
Training sample is constituted；
By all sample of misclassification and other off-line training samples in first Weak Classifier and second Weak Classifier
The 3rd training sample is constituted, the 3rd Weak Classifier is obtained by the study to the 3rd training sample, three are formed
The sorter network that two classification device is constituted, the 3rd training sample is made up of N number of off-line training sample.
Further, the step S2, be specially：
Image block is classified by the grader in the sorter network, the figure after being handled through each grader is obtained
As block；
The similarity between image block after calculating described image block and the processing through each grader, obtains multiple phases
Like degree；
According to the multiple similarity, the voting results of each grader are calculated；
According to the voting results, classification results are obtained, the classification results are the maximum class of multiple cumulative voting result.
Further, the size of described image block is at least one in 3*3,5*5,7*7 and 9*9.
Further, in described image block fault body image block, tomography edge image block and tomography background image block
Selection ratio is 3：2：1.
Second aspect, the invention provides a kind of Acceleration Algorithm in Seismic Coherence Cube image slices automatic identification based on AdaBoost algorithms
System, including：
Image block acquisition module, for obtaining the image block in Acceleration Algorithm in Seismic Coherence Cube faultage image, described image block includes disconnected
Layer body image block, tomography edge image block and tomography background image block；
Image block classification module, for the sorter network by training in advance, classifies to image block, obtains classification knot
Really, the sorter network is made up of multiple graders, and the sorter network is obtained by AdaBoost Algorithm Learnings；
Faultage image automatic identification module, for according to the classification results, realizing to the Acceleration Algorithm in Seismic Coherence Cube tomograph
The automatic identification of picture.
The Acceleration Algorithm in Seismic Coherence Cube image slices automatic recognition system based on AdaBoost algorithms that the present invention is provided, its technical side
Case is：Image block acquisition module is first passed through, for obtaining the image block in Acceleration Algorithm in Seismic Coherence Cube faultage image, described image block includes
Fault body image block, tomography edge image block and tomography background image block；
Then by image block classification module, for the sorter network by training in advance, image block is classified, obtained
To classification results, the sorter network is made up of multiple graders, and the sorter network is obtained by AdaBoost Algorithm Learnings；
Finally by faultage image automatic identification module, for according to the classification results, realizing to the Acceleration Algorithm in Seismic Coherence Cube tomograph
The automatic identification of picture.
The Acceleration Algorithm in Seismic Coherence Cube image slices automatic recognition system based on AdaBoost algorithms of the present invention, employs integrated
The method training sorter network of habit, classifies with sorter network to image block, can make automatic identification and the extraction of tomography
As a result it is more accurate.
Further, described image block sort module, specifically for training sorter network, be specially：
The first training sample of N number of off-line training sample formation is chosen, the study to first training sample obtains first
Individual Weak Classifier, the off-line training sample is the image block in Acceleration Algorithm in Seismic Coherence Cube faultage image；
The sample of misclassification in first Weak Classifier and other off-line training samples are constituted into second training sample,
Second Weak Classifier is obtained by the study to second training sample, second training sample is by N number of offline
Training sample is constituted；
By all sample of misclassification and other off-line training samples in first Weak Classifier and second Weak Classifier
The 3rd training sample is constituted, the 3rd Weak Classifier is obtained by the study to the 3rd training sample, three are formed
The sorter network that two classification device is constituted, the 3rd training sample is made up of N number of off-line training sample.
Further, described image block sort module, specifically for：
Image block is classified by the grader in the sorter network, the figure after being handled through each grader is obtained
As block；
The similarity between image block after calculating described image block and the processing through each grader, obtains multiple phases
Like degree；
According to the multiple similarity, the voting results of each grader are calculated；
According to the voting results, classification results are obtained, the classification results are the maximum class of multiple cumulative voting result.
Further, the size of described image block is at least one in 3*3,5*5,7*7 and 9*9.
Further, in described image block fault body image block, tomography edge image block and tomography background image block
Selection ratio is 3：2：1.
Brief description of the drawings
, below will be to specific in order to illustrate more clearly of the specific embodiment of the invention or technical scheme of the prior art
The accompanying drawing to be used needed for embodiment or description of the prior art is briefly described.
Fig. 1 shows a kind of Acceleration Algorithm in Seismic Coherence Cube image slices based on AdaBoost algorithms that the embodiment of the present invention is provided
The flow chart of automatic identifying method；
Fig. 2 shows a kind of Acceleration Algorithm in Seismic Coherence Cube image slices based on AdaBoost algorithms that the embodiment of the present invention is provided
The multicategory classification network diagram of automatic identifying method；
Fig. 3 shows a kind of Acceleration Algorithm in Seismic Coherence Cube image slices based on AdaBoost algorithms that the embodiment of the present invention is provided
The schematic diagram of automatic recognition system.
Embodiment
The embodiment of technical solution of the present invention is described in detail below in conjunction with accompanying drawing.Following examples are only used for
Clearly illustrate technical scheme, therefore be intended only as example, and the protection of the present invention can not be limited with this
Scope.
Embodiment one
Fig. 1 shows a kind of Acceleration Algorithm in Seismic Coherence Cube image slices based on AdaBoost algorithms that the embodiment of the present invention is provided
The flow chart of automatic identifying method；As shown in figure 1, a kind of Acceleration Algorithm in Seismic Coherence Cube based on AdaBoost algorithms that embodiment one is provided
Image slices automatic identifying method, including：
Step S1, obtains the image block in Acceleration Algorithm in Seismic Coherence Cube faultage image, and image block includes fault body image block, tomography side
Edge image block and tomography background image block；
Step S2, by the sorter network of training in advance, classifies to image block, obtains classification results, sorter network
It is made up of multiple graders, sorter network is obtained by AdaBoost Algorithm Learnings；
Step S3, according to classification results, realizes the automatic identification to Acceleration Algorithm in Seismic Coherence Cube faultage image.
The Acceleration Algorithm in Seismic Coherence Cube image slices automatic identifying method based on AdaBoost algorithms of the present invention, its technical scheme
It is：The image block in Acceleration Algorithm in Seismic Coherence Cube faultage image is obtained, image block includes fault body image block, tomography edge image block and broken
Layer background image block；By the sorter network of training in advance, image block is classified, classification results are obtained, sorter network by
Multiple graders are constituted, and sorter network is obtained by AdaBoost Algorithm Learnings；According to classification results, realize to Acceleration Algorithm in Seismic Coherence Cube
The automatic identification of faultage image.
The Acceleration Algorithm in Seismic Coherence Cube image slices automatic identifying method based on AdaBoost algorithms of the present invention, employs integrated
The method training sorter network of habit, Adaboost is a kind of iterative algorithm, and its core concept is to be directed to same Acceleration Algorithm in Seismic Coherence Cube
The training set of image formation, trains different graders (Weak Classifier), then these weak classifier sets is got up, and constitutes one
Individual sorter network, the classification capacity of sorter network is stronger.
By sorter network, change point of irregular tomography distribution, the complex-shaped changeable, tomography of tomography and background data
Cloth, it is whether correct according to the classification of each sample in each training set, and the general classification of last time accuracy rate, it is every to determine
The weights of individual sample.Sub-classification device is given by the new data set for changing weights to be trained, and finally obtains each training
Grader finally fusion get up, be used as last Decision Classfication device.Can exclude some using Adaboost graders need not
The training data feature wanted, and be placed on above the training data of key.By constantly training, the classification to data can be improved
Ability.The present invention trains obtained sorter network to classify image block with through Adaboost, can make the automatic knowledge of tomography
Result not with extraction is more accurate.
Referring to Fig. 2, the training process of sorter network is：
Choose N number of off-line training sample the first training sample of formation, the study of the first training sample is obtained first it is weak
Grader, off-line training sample is the image block in Acceleration Algorithm in Seismic Coherence Cube faultage image；
The sample of misclassification in first Weak Classifier and other off-line training samples are constituted into second training sample, passed through
Study to second training sample obtains second Weak Classifier, and second training sample is made up of N number of off-line training sample；
By all the sample of misclassification and other off-line training samples are constituted in first Weak Classifier and second Weak Classifier
3rd training sample, the 3rd Weak Classifier is obtained by the study to the 3rd training sample, forms three two classifications
The sorter network that device is constituted, the 3rd training sample is made up of N number of off-line training sample.
By constantly training, three two classification devices as shown in Figure 2, composition and classification network, to faultage image are obtained
Block is classified, and obtains more accurately classification results, and then make the automatic identification and the result of extraction of tomography more accurate.
Specifically, step S2 is specially：
Image block is classified by the grader in sorter network, the image after being handled through each grader is obtained
Block；
The similarity between image block after calculating image block and being handled through each grader, obtains multiple similarities；
Wherein, the detailed process of similarity between image block after calculating image block and being handled through each grader is：
In probability theory and statistical theory, Hellinger distances are used to measure the similarity of two probability distribution, therefore
Image block is calculated to the Hellinger distances of Different categories of samples the average image block gray scale Gauss Distribution Fitting to represent image block and warp
The similarity between image block after each grader processing, the computing formula of its Hellinger distance is：
WhereinFor current image block gray scale Gauss Distribution Fitting parameter；For the i-th class sample mean image
Block gray scale Gauss Distribution Fitting parameter.
According to multiple similarities, the voting results of each grader are calculated；
Calculated by below equation and obtain voting results：
Wherein vk={ vk(i) it is } the ballot value of k-th of two classification device；
According to voting results, classification results are obtained, classification results are the maximum class of multiple cumulative voting result.
Specifically, the size of image block is at least one in 3*3,5*5,7*7 and 9*9.
The minimum description yardstick that Acceleration Algorithm in Seismic Coherence Cube image interrupting layer edge pixel is found according to correlative study is 3*3, and is schemed
As the complete representation at edge generally requires 9*9 image block yardstick, therefore, have chosen respectively in the present invention 3*3,5*5,7*7 and
9*9 several yardsticks choose sample image block.
Preferably, understood through testing further comparing result, if image block is too small, such as 3*3 can then go out in classification results
The phenomenon that now continuous tomography is split off out, and if image block is excessive, such as 9*9, then occur in experimental result tomography edge and
Fault body has the effect of expansion.Only select moderate in tile size, in the case of such as 5*5 or 7*7, final classification knot
Fruit can just accurately identify fault body and tomography edge.Therefore the size of image block preferred in the present invention is 5*5 or 7*7.
Preferably, in image block the selection ratio of fault body image block, tomography edge image block and tomography background image block
Example is 3：2：1.
Consider the representativeness of sample data and the degree of rarefication problem of correspondence classification, i.e. the openness picture of background data to preferable,
And fault body image block, tomography edge image block and tomography background in the relatively complicated situation of tomography, image block of the present invention
The selection ratio of image block is 3：2：1.Fault body image block, tomography edge image block and tomography background image block can so be solved
Quantity is uneven, the problem of causing the automatic identification and the inaccurate result of extraction of tomography.
Fig. 3 shows a kind of Acceleration Algorithm in Seismic Coherence Cube image slices based on AdaBoost algorithms that the embodiment of the present invention is provided
The schematic diagram of automatic recognition system.As shown in figure 3, the invention provides a kind of Acceleration Algorithm in Seismic Coherence Cube figure based on AdaBoost algorithms
As tomography automatic recognition system 10, including：
Image block acquisition module 101, for obtaining the image block in Acceleration Algorithm in Seismic Coherence Cube faultage image, image block includes tomography
Body image block, tomography edge image block and tomography background image block；
Image block classification module 102, for the sorter network by training in advance, classifies to image block, is divided
Class result, sorter network is made up of multiple graders, and sorter network is obtained by AdaBoost Algorithm Learnings；
Faultage image automatic identification module 103, for according to classification results, realize to Acceleration Algorithm in Seismic Coherence Cube faultage image from
Dynamic identification.
The Acceleration Algorithm in Seismic Coherence Cube image slices automatic recognition system 10 based on AdaBoost algorithms that the present invention is provided, its technology
Scheme is：Image block acquisition module is first passed through, for obtaining the image block in Acceleration Algorithm in Seismic Coherence Cube faultage image, image block includes disconnected
Layer body image block, tomography edge image block and tomography background image block；
Then by image block classification module, for the sorter network by training in advance, image block is classified, obtained
To classification results, sorter network is made up of multiple graders, and sorter network is obtained by AdaBoost Algorithm Learnings；Finally by
Faultage image automatic identification module, for according to classification results, realizing the automatic identification to Acceleration Algorithm in Seismic Coherence Cube faultage image.
The Acceleration Algorithm in Seismic Coherence Cube image slices automatic recognition system based on AdaBoost algorithms of the present invention, employs integrated
The method training sorter network of habit, Adaboost is a kind of iterative algorithm, and its core concept is to be directed to same Acceleration Algorithm in Seismic Coherence Cube
The training set of image formation, trains different graders (Weak Classifier), then these weak classifier sets is got up, and constitutes one
Individual sorter network, the classification capacity of sorter network is stronger.
By sorter network, change point of irregular tomography distribution, the complex-shaped changeable, tomography of tomography and background data
Cloth, it is whether correct according to the classification of each sample in each training set, and the general classification of last time accuracy rate, it is every to determine
The weights of individual sample.Sub-classification device is given by the new data set for changing weights to be trained, and finally obtains each training
Grader finally fusion get up, be used as last Decision Classfication device.Can exclude some using Adaboost graders need not
The training data feature wanted, and be placed on above the training data of key.By constantly training, the classification to data can be improved
Ability.The present invention trains obtained sorter network to classify image block with through Adaboost, can make the automatic knowledge of tomography
Result not with extraction is more accurate.
Referring to Fig. 2, image block classification module 102, specifically for training sorter network, is specially：
Choose N number of off-line training sample the first training sample of formation, the study of the first training sample is obtained first it is weak
Grader, off-line training sample is the image block in Acceleration Algorithm in Seismic Coherence Cube faultage image；
The sample of misclassification in first Weak Classifier and other off-line training samples are constituted into second training sample, passed through
Study to second training sample obtains second Weak Classifier, and second training sample is made up of N number of off-line training sample；
By all the sample of misclassification and other off-line training samples are constituted in first Weak Classifier and second Weak Classifier
3rd training sample, the 3rd Weak Classifier is obtained by the study to the 3rd training sample, forms three two classifications
The sorter network that device is constituted, the 3rd training sample is made up of N number of off-line training sample.
By constantly training, three two classification devices as shown in Figure 2, composition and classification network, to faultage image are obtained
Block is classified, and obtains more accurately classification results, and then make the automatic identification and the result of extraction of tomography more accurate.
Specifically, image block classification module 102, specifically for：
Image block is classified by the grader in sorter network, the image after being handled through each grader is obtained
Block；
The similarity between image block after calculating image block and being handled through each grader, obtains multiple similarities；
Wherein, the detailed process of similarity between image block after calculating image block and being handled through each grader is：
In probability theory and statistical theory, Hellinger distances are used to measure the similarity of two probability distribution, therefore
Image block is calculated to the Hellinger distances of Different categories of samples the average image block gray scale Gauss Distribution Fitting to represent image block and warp
The similarity between image block after each grader processing, the computing formula of its Hellinger distance is：
WhereinFor current image block gray scale Gauss Distribution Fitting parameter；For the i-th class sample mean image
Block gray scale Gauss Distribution Fitting parameter.
According to multiple similarities, the voting results of each grader are calculated；
Calculated by below equation and obtain voting results：
Wherein vk={ vk(i) it is } the ballot value of k-th of two classification device；
According to voting results, classification results are obtained, classification results are the maximum class of multiple cumulative voting result.
Specifically, the size of image block is at least one in 3*3,5*5,7*7 and 9*9.
The minimum description yardstick that Acceleration Algorithm in Seismic Coherence Cube image interrupting layer edge pixel is found according to correlative study is 3*3, and is schemed
As the complete representation at edge generally requires 9*9 image block yardstick, therefore, have chosen respectively in the present invention 3*3,5*5,7*7 and
9*9 several yardsticks choose sample image block.
Preferably, understood through testing further comparing result, if image block is too small, such as 3*3 can then go out in classification results
The phenomenon that now continuous tomography is split off out, and if image block is excessive, such as 9*9, then occur in experimental result tomography edge and
Fault body has the effect of expansion.Only select moderate in tile size, in the case of such as 5*5 or 7*7, final classification knot
Fruit can just accurately identify fault body and tomography edge.Therefore the size of image block preferred in the present invention is 5*5 or 7*7.
Preferably, in image block the selection ratio of fault body image block, tomography edge image block and tomography background image block
Example is 3：2：1.
Consider the representativeness of sample data and the degree of rarefication problem of correspondence classification, i.e. the openness picture of background data to preferable,
And fault body image block, tomography edge image block and tomography background in the relatively complicated situation of tomography, image block of the present invention
The selection ratio of image block is 3：2：1.Fault body image block, tomography edge image block and tomography background image block can so be solved
Quantity is uneven, the problem of causing the automatic identification and the inaccurate result of extraction of tomography.
Finally it should be noted that：Various embodiments above is merely illustrative of the technical solution of the present invention, rather than its limitations；To the greatest extent
The present invention is described in detail with reference to foregoing embodiments for pipe, it will be understood by those within the art that：Its according to
The technical scheme described in foregoing embodiments can so be modified, or which part or all technical characteristic are entered
Row equivalent；And these modifications or replacement, the essence of appropriate technical solution is departed from various embodiments of the present invention technology
The scope of scheme, it all should cover among the claim of the present invention and the scope of specification.
Claims (10)
1. the Acceleration Algorithm in Seismic Coherence Cube image slices automatic identifying method based on AdaBoost algorithms, it is characterised in that including：
Step S1, obtains the image block in Acceleration Algorithm in Seismic Coherence Cube faultage image, and described image block includes fault body image block, tomography side
Edge image block and tomography background image block；
Step S2, by the sorter network of training in advance, classifies to image block, obtains classification results, the sorter network
It is made up of multiple graders, the sorter network is obtained by AdaBoost Algorithm Learnings；
Step S3, according to the classification results, realizes the automatic identification to the Acceleration Algorithm in Seismic Coherence Cube faultage image.
2. the Acceleration Algorithm in Seismic Coherence Cube image slices automatic identifying method according to claim 1 based on AdaBoost algorithms, its
It is characterised by,
The training process of the sorter network is：
Choose N number of off-line training sample the first training sample of formation, the study of first training sample is obtained first it is weak
Grader, the off-line training sample is the image block in Acceleration Algorithm in Seismic Coherence Cube faultage image；
The sample of misclassification in first Weak Classifier and other off-line training samples are constituted into second training sample, passed through
Study to second training sample obtains second Weak Classifier, and second training sample is by N number of off-line training
Sample is constituted；
By all the sample of misclassification and other off-line training samples are constituted in first Weak Classifier and second Weak Classifier
3rd training sample, the 3rd Weak Classifier is obtained by the study to the 3rd training sample, forms three two classes
The sorter network that grader is constituted, the 3rd training sample is made up of N number of off-line training sample.
3. the Acceleration Algorithm in Seismic Coherence Cube image slices automatic identifying method according to claim 1 based on AdaBoost algorithms, its
It is characterised by,
The step S2, be specially：
Image block is classified by the grader in the sorter network, the image after being handled through each grader is obtained
Block；
The similarity between image block after calculating described image block and the processing through each grader, obtains multiple similar
Degree；
According to the multiple similarity, the voting results of each grader are calculated；
According to the voting results, classification results are obtained, the classification results are the maximum class of multiple cumulative voting result.
4. the Acceleration Algorithm in Seismic Coherence Cube image slices automatic identifying method according to claim 1 based on AdaBoost algorithms, its
It is characterised by,
The size of described image block is at least one in 3*3,5*5,7*7 and 9*9.
5. the Acceleration Algorithm in Seismic Coherence Cube image slices automatic identifying method according to claim 1 based on AdaBoost algorithms, its
It is characterised by,
The selection ratio of fault body image block, tomography edge image block and tomography background image block in described image block is 3：2：
1。
6. the Acceleration Algorithm in Seismic Coherence Cube image slices automatic recognition system based on AdaBoost algorithms, it is characterised in that including：
Image block acquisition module, for obtaining the image block in Acceleration Algorithm in Seismic Coherence Cube faultage image, described image block includes fault body
Image block, tomography edge image block and tomography background image block；
Image block classification module, for the sorter network by training in advance, classifies to image block, obtains classification results,
The sorter network is made up of multiple graders, and the sorter network is obtained by AdaBoost Algorithm Learnings；
Faultage image automatic identification module, for according to the classification results, realizing to the Acceleration Algorithm in Seismic Coherence Cube faultage image
Automatic identification.
7. the Acceleration Algorithm in Seismic Coherence Cube image slices automatic recognition system according to claim 6 based on AdaBoost algorithms, its
It is characterised by,
Described image block sort module, specifically for training sorter network, be specially：
Choose N number of off-line training sample the first training sample of formation, the study of first training sample is obtained first it is weak
Grader, the off-line training sample is the image block in Acceleration Algorithm in Seismic Coherence Cube faultage image；
The sample of misclassification in first Weak Classifier and other off-line training samples are constituted into second training sample, passed through
Study to second training sample obtains second Weak Classifier, and second training sample is by N number of off-line training
Sample is constituted；
By all the sample of misclassification and other off-line training samples are constituted in first Weak Classifier and second Weak Classifier
3rd training sample, the 3rd Weak Classifier is obtained by the study to the 3rd training sample, forms three two classes
The sorter network that grader is constituted, the 3rd training sample is made up of N number of off-line training sample.
8. the Acceleration Algorithm in Seismic Coherence Cube image slices automatic recognition system according to claim 6 based on AdaBoost algorithms, its
It is characterised by,
Described image block sort module, specifically for：
Image block is classified by the grader in the sorter network, the image after being handled through each grader is obtained
Block；
The similarity between image block after calculating described image block and the processing through each grader, obtains multiple similar
Degree；
According to the multiple similarity, the voting results of each grader are calculated；
According to the voting results, classification results are obtained, the classification results are the maximum class of multiple cumulative voting result.
9. the Acceleration Algorithm in Seismic Coherence Cube image slices automatic recognition system according to claim 6 based on AdaBoost algorithms, its
It is characterised by,
The size of described image block is at least one in 3*3,5*5,7*7 and 9*9.
10. the Acceleration Algorithm in Seismic Coherence Cube image slices automatic recognition system according to claim 6 based on AdaBoost algorithms, its
It is characterised by,
The selection ratio of fault body image block, tomography edge image block and tomography background image block in described image block is 3：2：
1。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710186136.XA CN106951924B (en) | 2017-03-27 | 2017-03-27 | Seismic coherence body image fault automatic identification method and system based on AdaBoost algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710186136.XA CN106951924B (en) | 2017-03-27 | 2017-03-27 | Seismic coherence body image fault automatic identification method and system based on AdaBoost algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106951924A true CN106951924A (en) | 2017-07-14 |
| CN106951924B CN106951924B (en) | 2020-01-07 |
Family
ID=59473783
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710186136.XA Active CN106951924B (en) | 2017-03-27 | 2017-03-27 | Seismic coherence body image fault automatic identification method and system based on AdaBoost algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106951924B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109086773A (en) * | 2018-08-29 | 2018-12-25 | 电子科技大学 | Fault plane recognition methods based on full convolutional neural networks |
| CN111582359A (en) * | 2020-04-28 | 2020-08-25 | 新疆维吾尔自治区烟草公司 | Image identification method and device, electronic equipment and medium |
| CN113130018A (en) * | 2021-04-20 | 2021-07-16 | 东北石油大学 | Lithology identification method based on reservoir element target invariant feature description |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6018498A (en) * | 1998-09-02 | 2000-01-25 | Phillips Petroleum Company | Automated seismic fault detection and picking |
| CN103412331A (en) * | 2013-08-30 | 2013-11-27 | 电子科技大学 | Automatic extraction method for three-dimensional earthquake fault |
| CN103592681A (en) * | 2013-09-16 | 2014-02-19 | 电子科技大学 | Signal classification based seismic image horizon tracking method |
| CN105093289A (en) * | 2014-05-22 | 2015-11-25 | 中国石油化工股份有限公司 | Seismic horizon identification method based on image processing |
| CN105659292A (en) * | 2013-11-26 | 2016-06-08 | 界标制图有限公司 | Global grid building unfaulting sequence for complex fault-network topologies |
| CN106154327A (en) * | 2016-08-17 | 2016-11-23 | 中国石油化工股份有限公司 | A kind of method improving hidden fault recognizing precision |
-
2017
- 2017-03-27 CN CN201710186136.XA patent/CN106951924B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6018498A (en) * | 1998-09-02 | 2000-01-25 | Phillips Petroleum Company | Automated seismic fault detection and picking |
| CN103412331A (en) * | 2013-08-30 | 2013-11-27 | 电子科技大学 | Automatic extraction method for three-dimensional earthquake fault |
| CN103592681A (en) * | 2013-09-16 | 2014-02-19 | 电子科技大学 | Signal classification based seismic image horizon tracking method |
| CN105659292A (en) * | 2013-11-26 | 2016-06-08 | 界标制图有限公司 | Global grid building unfaulting sequence for complex fault-network topologies |
| CN105093289A (en) * | 2014-05-22 | 2015-11-25 | 中国石油化工股份有限公司 | Seismic horizon identification method based on image processing |
| CN106154327A (en) * | 2016-08-17 | 2016-11-23 | 中国石油化工股份有限公司 | A kind of method improving hidden fault recognizing precision |
Non-Patent Citations (5)
| Title |
|---|
| KRISTOFER M 等: "Semi-automatic detection of faults in 3D seismic data", 《GEOPHYSICAL PROSPECTING》 * |
| MAURICIO ARAYA-POLO: "Automated fault detection without seismic processing", 《THE LEADING EDGE》 * |
| 孙媛 等: "藏汉跨语言话题模型构建与对齐方法研究", 《中文信息学报》 * |
| 李转: "基于三维地震数据相干信息的断层识别方法研究", 《中国优秀硕士学位论文全文数据库 基础科学辑》 * |
| 许剑 等: "Adaboost算法分类器设计及其应用", 《四川理工学院学报（自然科学版）》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109086773A (en) * | 2018-08-29 | 2018-12-25 | 电子科技大学 | Fault plane recognition methods based on full convolutional neural networks |
| CN109086773B (en) * | 2018-08-29 | 2022-03-04 | 电子科技大学 | Fault plane identification method based on full convolution neural network |
| CN111582359A (en) * | 2020-04-28 | 2020-08-25 | 新疆维吾尔自治区烟草公司 | Image identification method and device, electronic equipment and medium |
| CN113130018A (en) * | 2021-04-20 | 2021-07-16 | 东北石油大学 | Lithology identification method based on reservoir element target invariant feature description |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106951924B (en) | 2020-01-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107657279B (en) | Remote sensing target detection method based on small amount of samples | |
| CN105844279B (en) | SAR image change detection based on deep learning and SIFT feature | |
| CN109583489A (en) | Defect classifying identification method, device, computer equipment and storage medium | |
| CN106408030B (en) | SAR image classification method based on middle layer semantic attribute and convolutional neural networks | |
| US20160196467A1 (en) | Three-Dimensional Face Recognition Device Based on Three Dimensional Point Cloud and Three-Dimensional Face Recognition Method Based on Three-Dimensional Point Cloud | |
| CN107133955B (en) | A kind of collaboration conspicuousness detection method combined at many levels | |
| CN107784276B (en) | Microseismic event identification method and device | |
| CN105574550A (en) | Vehicle identification method and device | |
| CN104573744B (en) | Fine granulation classification identifies and the part of object positions and feature extracting method | |
| CN105160310A (en) | 3D (three-dimensional) convolutional neural network based human body behavior recognition method | |
| CN105608456A (en) | Multi-directional text detection method based on full convolution network | |
| CN104102929A (en) | Hyperspectral remote sensing data classification method based on deep learning | |
| CN104021396A (en) | Hyperspectral remote sensing data classification method based on ensemble learning | |
| CN108171119B (en) | SAR image change detection method based on residual error network | |
| CN117115147A (en) | Textile detection method and system based on machine vision | |
| CN105741267A (en) | Multi-source image change detection method based on clustering guided deep neural network classification | |
| CN103996057A (en) | Real-time handwritten digital recognition method based on multi-feature fusion | |
| CN103927528B (en) | Coal and rock recognition method based on close neighborhood pixel gray level joint distribution characteristics | |
| CN104834941A (en) | Offline handwriting recognition method of sparse autoencoder based on computer input | |
| CN104978569B (en) | A kind of increment face identification method based on rarefaction representation | |
| CN114091606A (en) | Tunnel blasting blast hole half-hole mark identification and damage flatness evaluation classification method | |
| CN106446922A (en) | Crowd abnormal behavior analysis method | |
| CN114776304B (en) | Method, device, equipment and medium for identifying abnormities of deep sea mineral products | |
| Kim et al. | Deep learning based automatic defect classification in through-silicon Via process: FA: Factory automation | |
| CN106951924A (en) | Acceleration Algorithm in Seismic Coherence Cube image slices automatic identifying method and system based on AdaBoost algorithms |
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 |