CN104199092A - Multi-level framework based three-dimensional full-horizon automatic tracking method - Google Patents
Multi-level framework based three-dimensional full-horizon automatic tracking method Download PDFInfo
- Publication number
- CN104199092A CN104199092A CN201410440323.2A CN201410440323A CN104199092A CN 104199092 A CN104199092 A CN 104199092A CN 201410440323 A CN201410440323 A CN 201410440323A CN 104199092 A CN104199092 A CN 104199092A
- Authority
- CN
- China
- Prior art keywords
- layer
- horizon
- point
- layer position
- site
- 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
Abstract
The invention discloses a multi-level framework based three-dimensional full-horizon automatic tracking method. The multi-level framework based three-dimensional full-horizon automatic tracking method comprises the following steps of step one, inputting seismic data; step two, preprocessing the input data and obtaining positions of all horizon extreme points in an area required to be tracked; step three, processing the extreme points through a horizon extreme point connection based horizon fragment generation algorithm and connecting the extreme points into a horizon block; step four, further connecting the horizon block generated in the step three through a horizon fragment merging based horizon plane fusion algorithm to generate into a large horizon; step five, correcting the horizon through an expert system and obtaining a final tracking result; step six, outputting the horizon obtained by full-horizon tracking. The multi-level framework based three-dimensional full-horizon automatic tracking method has the advantages of achieving accurate three-dimensional full-horizon tracking in a complex geologic structure, fully utilizing seismic data waveform characteristics, horizon position distribution characteristics and mutual relationships between horizon and improving the seismic interpretation work efficiency and effect.
Description
Technical field
The present invention relates to a kind of automatic formation trace method, particularly a kind of three-dimensional holostrome position method for automatic tracking based on multi-level framework.
Background technology
Along with expanding economy and scientific and technical progress, people increase day by day to the demand of the mineral resources such as oil and natural gas, although the appearance of new forms of energy has reduced the dependence of people to traditional resource, but the most important resource that oil and the natural people of remaining depend on for existence, petroleum resources are buried at underground several kms place, finding dark oil gas field of ensconcing underground several kms is not an easy thing, the topmost method of Detecting Oil And Gas Fields is seismic prospecting at present, seismic prospecting is artificial excitation's seismic event on the ground, seismic event returns to ground after different geologic structure reflections, received by ground receiving sensor, and seismic record is got off, the subordinate phase of seismic prospecting is the seismic signal processing processing of getting off to receiving sensor record according to actual seismic characteristic, remove noise, form 3-D seismics image, the last stage of seismic prospecting is seismic interpretation, be exactly according to earthquake graphical analysis and prediction geologic structure, judgement and identification of hydrocarbon Storage.
Seismic interpretation is the model and forecast process about underground structure and attribute, underground sedimentary deposit is divided into some levels according to different rock characters and formation age, the directly perceived reaction of this hierarchical structure in geological data is layer position, layer position explains it is a very important part in seismic interpretation always, and accurate, effective layer position explains there is very important supporting role for the analysis of subsurface reservoir.Traditional layer position explains it is by manual interpretation always, and manual interpretation workload is large, the length that expends time in, and manual interpretation is explained that personnel's subjective consciousness affects, verifiability is poor.Existing automatic tracing technology can only be issued in the fairly simple situation of stratigraphic structure the requirement of structure elucidation, can not explain complicated tectonic structure.
The identification of layer position and tracking are parts very important in seismic interpretation, tracing of horizons is mainly the lineups of following the trail of in seismic image, traditional tracing of horizons mainly relies on geologist's manual interpretation, manual interpretation workload is large, consuming time for a long time, and cannot avoid the impact of subjectivity, occurred in recent years a lot of automatic formation trace methods, these methods mainly contain based on correlation technique, based on artificial neural network, based on genetic algorithm and the tracing of horizons technology based on finite mixtures Gauss etc.First extract layer digit wave form feature of tracing of horizons technology based on correlation technique, carry out the similarity degree of presentation layer digit wave form by the correlativity of waveform character, correlation method thinks that the large waveform of related coefficient belongs to same layer position, experiment shows that correlation method can identify some obvious layer positions, but significantly defect is exactly that tracking layer position is out sufficiently complete.Tracing of horizons technology based on artificial neural network is converted into pattern recognition problem tracing of horizons problem, artificial neural network carrys out training network with a large amount of training samples, then by the training pattern obtaining, seismic image is carried out to tagsort, Artificial Neural Network can be identified obvious geologic feature, but the quality of following the trail of effect depends primarily on the selection of training sample set, the a large amount of manual intervention needing in training process, and it is longer to expend time in, and interpretation is poor.Tracing of horizons technology based on genetic algorithm is mainly that seismic image is carried out to model analysis, tracing of horizons problem is converted into constrained optimization problem, then solve by genetic algorithm, the tracing of horizons technology subject matter based on genetic algorithm is inadequate to the precision of approaching and solving of complicated geological structure.Tracing of horizons technology based on finite mixtures Gauss is converted into tracing of horizons problem the clustering problem of feature set, thinking that waveform character gathers belongs to same layer position in of a sort layer site, this method has the ability of good crossover fault, but needs a large amount of manual interventions.In a word, existing tracing of horizons technology is difficult to reach the requirement of seismic prospecting in effect and automaticity, and the present invention proposes the full automatic formation trace technology of a kind of new three-dimensional, and practical application shows the present invention's needs of seismic exploration contentedly.
Summary of the invention
The object of the invention is to overcome the deficiencies in the prior art, a kind of new method that proposes a kind of tracing of horizons is provided, realize in complicated geological structure accurately three-dimensional full tracing of horizons, take full advantage of waveform characteristic, layer position position distribution characteristic and the layer interdigit mutual relationship of geological data, improved the three-dimensional holostrome position method for automatic tracking based on multi-level framework of efficiency and the effect of seismic interpretation work.
The object of the invention is to be achieved through the following technical solutions: the three-dimensional holostrome position method for automatic tracking based on multi-level framework, comprises the following steps:
S1: input geological data;
S2: the data of inputting are carried out to pre-service, obtain the position of all layer position extreme points that need trace regions;
S3: utilize the layer bit slice section generating algorithm connecting based on layer position extreme point to process extreme point, extreme point is connected to stratification position piece;
S4: the layer position piece that uses the layer plane blending algorithm merging based on layer bit slice section to produce step S3 further connects, and generates large layer position;
S5: use expert system to revise layer position, finally followed the trail of result;
S6: export the layer position that full tracing of horizons obtains.
Further, in described step S2, carrying out the pretreated concrete grammar of data is: input geological data is discrete three-dimensional data, what in data volume, store is the amplitude of seismic waveshape, three dimensions of data volume are respectively inline, crossline and time, three dimension scopes are respectively inline section numbers, the number of crossline section number and time direction sampled point, a seismic waveshape when take out one geological data in time direction, geological data pre-service is exactly that each seismic waveshape of processing in 3-D data volume obtains all extreme points on seismic waveshape, the value of i point on seismic waveshape is designated as k
iif i point is maximum point, put the k that satisfies condition for i
i-1< k
i, k
i+1< k
i, utilize this relation to obtain layer position extreme points all in data volume.
Further, the layer bit slice section generating algorithm in described step S3 comprises following sub-step:
S31: read in all layers site;
S32: judge whether to have not processed layer site, if do not exist algorithm to finish, generate all layers position, get a not processed layer site otherwise appoint;
S33: judge whether this layer of site is initiatively to generate a little, this point is put into and treated traversal set if initiatively generate point, carry out next step operation, this point is labeled as to passive generation point if not initiatively generating point, return to step S32;
S34: judge and treat whether traversal set is empty, traversal set represents that for sky a layer position generates successfully if treat, layer position id adds 1, returns to step S32; If treat, traversal set is for empty, from treat traversal set, appoints to get any and carry out next step operation;
S35: reject the abnormity point in this epsilon neighborhood, add layer position id to the point in neighborhood, in traversal epsilon, institute a little, if initiatively generate a little, join and treat traversal set, if not initiatively generating a little, be labeled as passive generation point, return to step S34.
Described active generates point and is defined as: in the layer site of inside, layer position, the number in epsilon neighborhood internal layer site is greater than threshold value threshold.
Described passive generation point is defined as: in the layer site of layer position marginal position, the number in epsilon neighborhood internal layer site is less than threshold value threshold, in epsilon neighborhood, has and initiatively generates a little.
The described layer site that peel off is defined as: in the layer site of layer position marginal position, in epsilon neighborhood, initiatively do not generate a little.
Further, described step S4 layer plane blending algorithm comprises following sub-step:
S41: the layer position that input layer bit slice section generating algorithm produces;
S42: cluster is carried out in all layers site: first find out layer positions all in geological data, near waveform removing layer site, seismic waveshape is carried out to Chebyshev's matching and obtain the feature of fitting coefficient as waveform, obtain the characteristic set in layer site, carry out cluster for the characteristic set in layer site with gauss hybrid models algorithm, obtain cluster result;
S43: appoint and get a layer position;
S44: judge whether a layer position size reaches work area scope, if layer position size reach work area scope presentation layer position be complete, directly step S43 is also returned in output layer position, otherwise carries out next step operation;
S45: judged whether to gather in of a sort layer position, if there is no gathered the layer position in same class, expression does not have and this layer of layer position that digit wave form is similar, and layer position is without expansion, and step S43 is also returned in direct output layer position, otherwise carries out next step operation,
S46: whether judgement gathers has overlapping in of a sort layer position, because the layer position in 3D seismic data is the curved surface that has fluctuating, but in every one, only has a layer site in each layer of position, can not merge so have between overlapping layer position, overlapping if layer position has, step S43 is also returned in direct output layer position, otherwise carries out next step operation;
S47: merge two layer positions, return to step S44.
The invention has the beneficial effects as follows:
1, a kind of new method of tracing of horizons is proposed, realize in complicated geological structure accurately three-dimensional full tracing of horizons, take full advantage of waveform characteristic, layer position position distribution characteristic and the layer interdigit mutual relationship of geological data, can in complicated geologic structure, realize exactly full automatic formation trace, have outstanding performance in performance aspect the complicated seismotectonics of explanation, crossover fault and layer bit integrity, can effectively improve efficiency and the effect of seismic interpretation work;
2, can replace manual interpretation completely, without manual intervention, realize the full automatic formation trace in work area on a large scale.
Brief description of the drawings
Fig. 1 is the process flow diagram of automatic formation trace method of the present invention;
Fig. 2 is the waveform schematic diagram that extracts from seismic data volume of the present invention;
Fig. 3 extracts extreme point schematic diagram in a section of the present invention;
Fig. 4 is layer bit slice section generating algorithm process flow diagram of the present invention;
Fig. 5 is layer of the present invention site classification schematic diagram;
Fig. 6 is layer bit slice section generating algorithm schematic diagram of the present invention;
Fig. 7 is layer plane blending algorithm process flow diagram of the present invention;
Fig. 8 is the process flow diagram that in the present embodiment HFM algorithm, all layers site is carried out cluster;
Fig. 9 is that the layer position of the present embodiment merges schematic diagram;
Figure 10 is the HFM algorithm effect figure of the present embodiment;
Figure 11 is automatic formation trace result vertical view between certain work area H02 of the present embodiment and H023;
Figure 12 is full automatic formation trace result and practical logging explanation comparison diagram between certain work area H02 of the present embodiment and H023;
Figure 13 is the present embodiment F3 work area business software automatic tracing result;
Figure 14 is the F3 work area of the present embodiment automatic tracing result of the present invention.
Embodiment
1, position, 3D seismic data middle level location lookup method
Data for structure elucidation are 3-D seismics view data, 3-D seismics view data can be expressed as a function S=s (x who contains three independents variable, y, z), it is generally acknowledged that x, y, z represents respectively three direction inline directions, crossline direction and time directions in geological data.X=x
0time, S=s (x
0, y, z) and represent an inline section in geological data, same y=y
0time, S=s (x, y
0, z) represent a crossline section in geological data, work as x=x
0, y=y
0time, S=s (x
0, y
0, z) represent the track data in 3D seismic data, a namely seismic waveshape.For the 3D seismic data through noise reduction process, it is generally acknowledged that layer position is arranged in the extreme point position of seismic waveshape, so layer site (x
0, y
0, z
0) satisfy condition:
Layer site formation curved surfaces all in 3-D seismics view data can be expressed as:
Layer position location lookup method is exactly to utilize this character in layer site to find out layer sites all in 3-D seismics view data.
2, seismic wave characteristic extracting method in 3D seismic data
The 3D seismic data obtaining in engineering application and scientific experiment is generally discrete, seismic waveshape in 3D seismic data is some discrete points like this, first the waveform characteristic of Study of Seismic data will recover seismic waveshape, namely discrete point is carried out to approximation of function, approximation of function is mainly matching and interpolation, the method of interpolation is mainly used in data point more accurately in situation, the interpolating function obtaining is through all discrete data points, and approximating method is mainly the trend of matching discrete point, discrete data point needn't one be positioned on fitting function, 3D seismic data is generally more coarse data, the general method that uses matching is carried out approximation of function to seismic waveshape, different approximating methods is mainly the selection difference of fitting function, conventional fitting function has linear function, polynomial function, exponential function, trigonometric function or other function.Show good characteristic in Chebyshev's matching aspect seismic waveshape approximation of function, generally use chebyshev function to carry out matching to seismic waveshape.
Seismic wave characteristic is extracted and exactly the waveform after matching is carried out to feature extraction, conventional waveform character has the amplitude of waveform, the cycle of waveform, the energy of waveform etc., tracing of horizons is mainly the shape based on waveform, research finds that Chebyshev's fitting coefficient of seismic waveshape exists very strong statistical property, near the approximate Gaussian distributed of the same order fitting coefficient of waveform same layer position, the approximate Multidimensional and Hybrid Gaussian distribution of obeying of multistage fitting coefficient, Chebyshev's fitting coefficient of the waveform of different layers position is obeyed different Multidimensional and Hybrid Gaussian distribution, seismic wave characteristic can represent with Chebyshev's fitting coefficient like this.
3, seismic wave characteristic clustering method in 3D seismic data
Using Chebyshev's fitting coefficient in all layers site in geological data as input set, close and carry out cluster in input set, Chebyshev's fitting coefficient in layer site is polymerized to different bunches, corresponding layer site is just arranged in different classes, layer site can be divided into different classes by this method, in this method, clustering algorithm is only used the waveform character of layer position, does not use a layer bit position relation, as long as just gathering in same class in the layer site of waveform similarity.
The kind of clustering algorithm is a lot, mainly contains cluster, the cluster based on density, the cluster based on layering, the cluster based on grid and the cluster based on model etc. based on dividing.Selecting which type of clustering algorithm is mainly according to data characteristics and application requirements, in geological data, seismic wave characteristic is obeyed Multidimensional and Hybrid Gaussian distribution, GMM algorithm is the main method of carrying out limited gauss hybrid models maximal possibility estimation, has good Clustering Effect for the feature that meets Multidimensional and Hybrid Gaussian distribution.
Further illustrate technical scheme of the present invention below in conjunction with the drawings and specific embodiments, but the content that the present invention protects is not limited to the following stated.
As shown in Figure 1, the three-dimensional holostrome position method for automatic tracking based on multi-level framework, comprises the following steps:
S1: input geological data;
S2: the data of inputting are carried out to pre-service, obtain the position of all layer position extreme points that need trace regions;
S3: utilize the layer bit slice section generating algorithm connecting based on layer position extreme point to process extreme point, extreme point is connected to stratification position piece;
S4: the layer position piece that uses the layer plane blending algorithm merging based on layer bit slice section to produce step S3 further connects, and generates large layer position;
S5: use expert system to revise layer position, finally followed the trail of result;
S6: export the layer position that full tracing of horizons obtains.
Further, in described step S2, carrying out the pretreated concrete grammar of data is: input geological data is discrete three-dimensional data, what in data volume, store is the amplitude of seismic waveshape, three dimensions of data volume are respectively inline, crossline and time, three dimension scopes are respectively inline section numbers, the number of crossline section number and time direction sampled point is a seismic waveshape when take out one geological data in time direction, as shown in Figure 2.Geological data pre-service is exactly that each seismic waveshape of processing in 3-D data volume obtains all extreme points on seismic waveshape, and the value of i point on seismic waveshape is designated as k
iif i point is maximum point, put the k that satisfies condition for i
i-1< k
i, k
i+1< k
i, utilize this relation to obtain layer position extreme points all in data volume, as shown in Figure 3.
As shown in Figure 4, the layer bit slice section generating algorithm (HEC) in the step S3 described in the present embodiment comprises following sub-step:
S31: read in all layers site;
S32: judge whether to have not processed layer site, if do not exist algorithm to finish, generate all layers position, get a not processed layer site otherwise appoint;
S33: judge whether this layer of site is initiatively to generate a little, this point is put into and treated traversal set if initiatively generate point, carry out next step operation, this point is labeled as to passive generation point if not initiatively generating point, return to step S32; So as Fig. 3, the some D in figure initiatively generates a little, accessed mistake of the point on the D left side, next accessed point is E point, is exactly to generate point and passive generation point by the new active of so continuous generation, layer bit slice section is constantly expanded;
S34: judge and treat whether traversal set is empty, traversal set represents that for sky a layer position generates successfully if treat, layer position id adds 1, returns to step S32; If treat, traversal set is for empty, from treat traversal set, appoints to get any and carry out next step operation;
S35: reject the abnormity point in this epsilon neighborhood, add layer position id to the point in neighborhood, in traversal epsilon, institute a little, if initiatively generate a little, join and treat traversal set, if not initiatively generating a little, be labeled as passive generation point, return to step S34.As shown in Figure 6, point E in Fig. 6 is one and initiatively generates a little, what inner ring showed is the epsilon neighborhood of E, outer ring is the epsilon neighborhood of 2 times, outer ring is for rejecting abnormalities point, in the epsilon of E neighborhood, in neighborhood, only having a some G is the point of not accessing, but G and E not necessarily belong to same layer position, if G and E do not belong to same layer position, G is the abnormity point in the epsilon neighborhood of E, the 2epsilon neighborhood of E is for rejecting the abnormity point in epsilon neighborhood, as can be seen from the figure in lower one, the 2epsilon neighborhood of E has included a G and H, like this by G point, the waveform at H point and layer E place, site is asked related coefficient, retain the layer site of a waveform the most similar with the waveform at E place, if G and E are not same layer positions like this, and H and E are the words of same layer position, the related coefficient of H point and E point waveform can be greater than the waveform correlation coefficient of E point and G, will reject like this G point and retain H point, thereby reach the effect of rejecting abnormalities point.
Described active generates point and is defined as: in the layer site of inside, layer position, the number in epsilon neighborhood internal layer site is greater than threshold value threshold.As shown in Fig. 5 mid point A, it is 6 that threshold value threshold is made as the number of putting in the epsilon neighborhood of 3, A, is greater than threshold value threshold, and A is for initiatively to generate a little.
Described passive generation point is defined as: in the layer site of layer position marginal position, the number in epsilon neighborhood internal layer site is less than threshold value threshold, in epsilon neighborhood, has and initiatively generates a little.As shown in Fig. 5 mid point C, in C point epsilon neighborhood, the number of point is 2, is less than threshold value threshold, so C initiatively generates a little, initiatively generates a little because the epsilon neighborhood of C contains, and C is passive generation point.
The described layer site that peel off is defined as: in the layer site of layer position marginal position, in epsilon neighborhood, initiatively do not generate a little.As shown in Fig. 5 mid point B, in B point epsilon neighborhood, the number of point is 1, is less than threshold value threshold, so B initiatively generates a little, does not contain and initiatively generate a little in the epsilon neighborhood due to B, and B is layer site that peel off.
Little tomography can only be crossed in the layer position that HEC algorithm obtains, and when the layer distance of positions of tomography is when being greater than epsilon, HEC algorithm cannot be cross-domain, and a layer position can form two; In addition, in the time of the poor continuity of layer position, HEC algorithm can form layer bit slice section one by one, cannot be linked to be large layer position.The layer bit slice section that HFM algorithm produces HEC algorithm connects, and the layer bit slice section that belongs to same layer position can be connected into large layer position, can successfully solve these two defects of HEC algorithm.As shown in Figure 7, the step S4 layer plane blending algorithm (HFM) described in the present embodiment comprises following sub-step:
S41: the layer position that input layer bit slice section generating algorithm produces;
S42: cluster is carried out in all layers site: first find out layer positions all in geological data, near waveform removing layer site, seismic waveshape is carried out to Chebyshev's matching and obtain the feature of fitting coefficient as waveform, obtain the characteristic set in layer site, carry out cluster for the characteristic set in layer site with gauss hybrid models algorithm, obtain cluster result, its idiographic flow as shown in Figure 8;
S43: appoint and get a layer position;
S44: judge whether a layer position size reaches work area scope, if layer position size reach work area scope presentation layer position be complete, directly step S43 is also returned in output layer position, otherwise carries out next step operation;
S45: judged whether to gather in of a sort layer position, if there is no gathered the layer position in same class, expression does not have and this layer of layer position that digit wave form is similar, and layer position is without expansion, and step S43 is also returned in direct output layer position, otherwise carries out next step operation,
S46: whether judgement gathers has overlapping in of a sort layer position, because the layer position in 3D seismic data is the curved surface that has fluctuating, but in every one, only has a layer site in each layer of position, can not merge so have between overlapping layer position, overlapping if layer position has, step S43 is also returned in direct output layer position, otherwise carries out next step operation;
S47: merge two layer positions, return to step S44; As shown in Figure 9, A is to be connected by HEC algorithm the layer bit slice section generating with B, belong in same class in gauss hybrid models algorithm (GMM algorithm) position, middle level Segment A and B, and A and B do not have layer position overlapping, so layer position A and B are connected into a layer position in HFM algorithm.
HFM algorithm effect figure of the present invention as shown in figure 10, in figure, a) be the layer bit slice section that HEC algorithm produces, can find out that a lot of layer bit slice sections is not complete layer position, b) be the result figure of GMM algorithm cluster, from b), can find out that some layer of bit slice section gathered in same class, be c) design sketch after FLG merges result a) and b), and a lot of layers of bit slice section in can finding out a) have been fused to same layer position.
Because actual geologic structure is very complicated, in seismic prospecting and geologic structure interpretation work, there is the phenomenon that cannot explain with existing geological theory, to adjust according to actual exploration result the structure elucidation in earthquake work area, the layer position modified result algorithm of the expert's level in this algorithm flow is just based on these actual conditions, allow veteran structure elucidation expert in conjunction with reality exploration result, full tracing of horizons result be revised, make to follow the trail of the more realistic tectonic structure of result.
The present invention applies in actual work area:
The present embodiment has been chosen sub-section on work area, southwest palpus two and has been carried out full automatic formation trace work, follows the trail of the weak reflection line-ups at objective interval top, i.e. mound shape, groove shape seismic facies seismic facies, and then study advantageously seismic phase spatial distribution characteristic.Figure 11 carries out the layer position fragment vertical view that full tracing of horizons obtains between H03 and H23, in order to verify the correctness of the holostrome position method for tracing that the present invention proposes, full tracing of horizons result between H03 of the present invention and H23 and actual well logging result are contrasted, as shown in figure 12.As can be seen from the figure, because the layer position between two-layer is too broken little, the method of manual interpretation is difficult to identify all layer position fragments, and in Figure 12, can be clear that the holostrome position method for automatic tracking that the present invention proposes can track out all destination layer positions resolutely exactly, the full tracing of horizons result that the present invention proposes and real logging data have well identical.
The present invention program and business software contrast: F3 work area is the block that the North Sea is positioned at Holland's part, this block has been done to 3D earthquake-capturing, object is to carry out upper Jurassic systerm---the oil-gas exploration of lower Cretaceous strata, due to F3 work area complicated geological configuration, it is the difficult point of structure elucidation always, the present invention selects this work area testing algorithm performance, and contrast with the current best business software of performance, Figure 13 is F3 work area business software automatic tracing result, and Figure 14 is work area automatic tracing result of the present invention.Meticulousr from following the trail of the layer position that result can find out that the present invention follows the trail of out, there are a lot of layers of position business software not follow the trail of out, out tracked in the present invention.
In order to test algorithm time efficiency of the present invention, select the data of different range size in certain work area to carry out the trace test of holostrome position, test result as shown in Table 1:
Table one
| Work area scope | Time |
| 100*100*30 | 17s |
| 200*200*30 | 68s |
| 400*400*30 | 260s |
The computer that uses configuration as table and as shown in:
Table two
| Processor | Intel(R)Core(TM)i3-3240 |
| CPU | 3.40GHz |
| RAM | 4.00G (3.71G can use) |
| Operating system | Windows7 |
| Hard disk | 500GB |
Those of ordinary skill in the art will appreciate that, embodiment described here is in order to help reader understanding's principle of the present invention, should be understood to that protection scope of the present invention is not limited to such special statement and embodiment.Those of ordinary skill in the art can make various other various concrete distortion and combinations that do not depart from essence of the present invention according to these technology enlightenments disclosed by the invention, and these distortion and combination are still in protection scope of the present invention.
Claims (7)
1. the three-dimensional holostrome position method for automatic tracking based on multi-level framework, is characterized in that, comprises the following steps:
S1: input geological data;
S2: the data of inputting are carried out to pre-service, obtain the position of all layer position extreme points that need trace regions;
S3: utilize the layer bit slice section generating algorithm connecting based on layer position extreme point to process extreme point, extreme point is connected to stratification position piece;
S4: the layer position piece that uses the layer plane blending algorithm merging based on layer bit slice section to produce step S3 further connects, and generates large layer position;
S5: use expert system to revise layer position, finally followed the trail of result;
S6: export the layer position that full tracing of horizons obtains.
2. three-dimensional holostrome according to claim 1 position method for automatic tracking, it is characterized in that, in described step S2, carrying out the pretreated concrete grammar of data is: input geological data is discrete three-dimensional data, what in data volume, store is the amplitude of seismic waveshape, three dimensions of data volume are respectively inline, crossline and time, three dimension scopes are respectively inline section numbers, the number of crossline section number and time direction sampled point, a seismic waveshape when take out one geological data in time direction, geological data pre-service is exactly that each seismic waveshape of processing in 3-D data volume obtains all extreme points on seismic waveshape, the value of i point on seismic waveshape is designated as k
iif i point is maximum point, put the k that satisfies condition for i
i-1< k
i, k
i+1< k
i, utilize this relation to obtain layer position extreme points all in data volume.
3. three-dimensional holostrome according to claim 1 position method for automatic tracking, is characterized in that, the layer bit slice section generating algorithm in described step S3 comprises following sub-step:
S31: read in all layers site;
S32: judge whether to have not processed layer site, if do not exist algorithm to finish, generate all layers position, get a not processed layer site otherwise appoint;
S33: judge whether this layer of site is initiatively to generate a little, this point is put into and treated traversal set if initiatively generate point, carry out next step operation, this point is labeled as to passive generation point if not initiatively generating point, return to step S32;
S34: judge and treat whether traversal set is empty, traversal set represents that for sky a layer position generates successfully if treat, layer position id adds 1, returns to step S32; If treat, traversal set is for empty, from treat traversal set, appoints to get any and carry out next step operation;
S35: reject the abnormity point in this epsilon neighborhood, add layer position id to the point in neighborhood, in traversal epsilon, institute a little, if initiatively generate a little, join and treat traversal set, if not initiatively generating a little, be labeled as passive generation point, return to step S34.
4. three-dimensional holostrome according to claim 3 position method for automatic tracking, is characterized in that, described active generates point and is defined as: in the layer site of inside, layer position, the number in epsilon neighborhood internal layer site is greater than threshold value threshold.
5. three-dimensional holostrome according to claim 3 position method for automatic tracking, it is characterized in that, described passive generation point is defined as: in the layer site of layer position marginal position, the number in epsilon neighborhood internal layer site is less than threshold value threshold, in epsilon neighborhood, has and initiatively generates a little.
6. three-dimensional holostrome according to claim 3 position method for automatic tracking, is characterized in that, the described layer site that peel off is defined as: in the layer site of layer position marginal position, in epsilon neighborhood, initiatively do not generate a little.
7. three-dimensional holostrome according to claim 3 position method for automatic tracking, is characterized in that, described step S4 layer plane blending algorithm comprises following sub-step:
S41: the layer position that input layer bit slice section generating algorithm produces;
S42: cluster is carried out in all layers site: first find out layer positions all in geological data, near waveform removing layer site, seismic waveshape is carried out to Chebyshev's matching and obtain the feature of fitting coefficient as waveform, obtain the characteristic set in layer site, carry out cluster for the characteristic set in layer site with gauss hybrid models algorithm, obtain cluster result;
S43: appoint and get a layer position;
S44: judge whether a layer position size reaches work area scope, if layer position size reach work area scope presentation layer position be complete, directly step S43 is also returned in output layer position, otherwise carries out next step operation;
S45: judged whether to gather in of a sort layer position, if there is no gathered the layer position in same class, expression does not have and this layer of layer position that digit wave form is similar, and layer position is without expansion, and step S43 is also returned in direct output layer position, otherwise carries out next step operation,
S46: whether judgement gathers has overlapping in of a sort layer position, because the layer position in 3D seismic data is the curved surface that has fluctuating, but in every one, only has a layer site in each layer of position, can not merge so have between overlapping layer position, overlapping if layer position has, step S43 is also returned in direct output layer position, otherwise carries out next step operation;
S47: merge two layer positions, return to step S44.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410440323.2A CN104199092A (en) | 2014-08-31 | 2014-08-31 | Multi-level framework based three-dimensional full-horizon automatic tracking method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410440323.2A CN104199092A (en) | 2014-08-31 | 2014-08-31 | Multi-level framework based three-dimensional full-horizon automatic tracking method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104199092A true CN104199092A (en) | 2014-12-10 |
Family
ID=52084404
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410440323.2A Pending CN104199092A (en) | 2014-08-31 | 2014-08-31 | Multi-level framework based three-dimensional full-horizon automatic tracking method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104199092A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105182410A (en) * | 2015-09-02 | 2015-12-23 | 电子科技大学 | Seismic data layered characteristic reinforcement drafting method |
| CN105223609A (en) * | 2015-09-18 | 2016-01-06 | 电子科技大学 | Based on the 3-D seismics image holostrome position method for automatic tracking of match search |
| CN109709603A (en) * | 2018-11-23 | 2019-05-03 | 中国石油天然气股份有限公司 | Seismic horizon identification and method for tracing, system |
| CN110031895A (en) * | 2019-03-11 | 2019-07-19 | 西安科技大学 | A kind of Multiple-Point Geostatistics stochastic inversion methods and device based on image stitching |
| CN111158046A (en) * | 2020-01-09 | 2020-05-15 | 中国科学技术大学 | Earthquake horizon automatic extraction device and method based on dynamic programming |
| CN111796324A (en) * | 2019-04-09 | 2020-10-20 | 中国石油天然气股份有限公司 | Seismic all-horizon tracking method and device |
| CN112180447A (en) * | 2019-07-04 | 2021-01-05 | 中国石油天然气集团有限公司 | Method and system for eliminating strong reflection shielding of reservoir |
| CN113687423A (en) * | 2021-08-04 | 2021-11-23 | 潜能恒信能源技术股份有限公司 | Seismic data interpretation method and system based on interface extraction |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1797039A (en) * | 2004-12-29 | 2006-07-05 | 中国石油天然气集团公司 | Method for automatic tracking 3D geological horizon |
| CN102819040A (en) * | 2012-08-22 | 2012-12-12 | 电子科技大学 | Three-dimensional seismic horizon automatic tracking method based on central dispersion and dip angle attribute |
| CN102819688A (en) * | 2012-08-29 | 2012-12-12 | 电子科技大学 | Two-dimensional seismic data full-layer tracking method based on semi-supervised classification |
| US20130083973A1 (en) * | 2011-10-04 | 2013-04-04 | Sebastien Guillon | Method of tracking discontinuous seismic horizons in seismic images |
| CN103901467A (en) * | 2014-03-18 | 2014-07-02 | 中国石油集团川庆钻探工程有限公司地球物理勘探公司 | Method for tracking positions of three-dimensional seismic data |
-
2014
- 2014-08-31 CN CN201410440323.2A patent/CN104199092A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1797039A (en) * | 2004-12-29 | 2006-07-05 | 中国石油天然气集团公司 | Method for automatic tracking 3D geological horizon |
| US20130083973A1 (en) * | 2011-10-04 | 2013-04-04 | Sebastien Guillon | Method of tracking discontinuous seismic horizons in seismic images |
| CN102819040A (en) * | 2012-08-22 | 2012-12-12 | 电子科技大学 | Three-dimensional seismic horizon automatic tracking method based on central dispersion and dip angle attribute |
| CN102819688A (en) * | 2012-08-29 | 2012-12-12 | 电子科技大学 | Two-dimensional seismic data full-layer tracking method based on semi-supervised classification |
| CN103901467A (en) * | 2014-03-18 | 2014-07-02 | 中国石油集团川庆钻探工程有限公司地球物理勘探公司 | Method for tracking positions of three-dimensional seismic data |
Non-Patent Citations (3)
| Title |
|---|
| 刘宝和: "《中国石油勘探开发百科全书·综合卷》", 31 December 2008, 石油工业出版社 * |
| 涂先见: "基于信号分类的地震层位追踪方法研究", 《中国优秀硕士学位论文全文数据库·基础科学辑》 * |
| 陈耿 等: "《审计知识工程》", 31 October 2006, 清华大学出版社 * |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105182410A (en) * | 2015-09-02 | 2015-12-23 | 电子科技大学 | Seismic data layered characteristic reinforcement drafting method |
| CN105223609A (en) * | 2015-09-18 | 2016-01-06 | 电子科技大学 | Based on the 3-D seismics image holostrome position method for automatic tracking of match search |
| CN105223609B (en) * | 2015-09-18 | 2017-09-19 | 电子科技大学 | 3-D seismics image holostrome position method for automatic tracking based on matching search |
| CN109709603A (en) * | 2018-11-23 | 2019-05-03 | 中国石油天然气股份有限公司 | Seismic horizon identification and method for tracing, system |
| CN110031895A (en) * | 2019-03-11 | 2019-07-19 | 西安科技大学 | A kind of Multiple-Point Geostatistics stochastic inversion methods and device based on image stitching |
| CN110031895B (en) * | 2019-03-11 | 2020-12-15 | 西安科技大学 | Multipoint geostatistical stochastic inversion method and device based on image stitching |
| CN111796324A (en) * | 2019-04-09 | 2020-10-20 | 中国石油天然气股份有限公司 | Seismic all-horizon tracking method and device |
| US11531130B2 (en) | 2019-04-09 | 2022-12-20 | Petrochina Company Limited | Seismic full horizon tracking method, computer device and computer-readable storage medium |
| CN111796324B (en) * | 2019-04-09 | 2023-02-10 | 中国石油天然气股份有限公司 | Seismic all-horizon tracking method and device |
| CN112180447A (en) * | 2019-07-04 | 2021-01-05 | 中国石油天然气集团有限公司 | Method and system for eliminating strong reflection shielding of reservoir |
| CN111158046A (en) * | 2020-01-09 | 2020-05-15 | 中国科学技术大学 | Earthquake horizon automatic extraction device and method based on dynamic programming |
| CN113687423A (en) * | 2021-08-04 | 2021-11-23 | 潜能恒信能源技术股份有限公司 | Seismic data interpretation method and system based on interface extraction |
| CN113687423B (en) * | 2021-08-04 | 2024-03-01 | 潜能恒信能源技术股份有限公司 | Seismic data interpretation method and system based on interface extraction |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104199092A (en) | Multi-level framework based three-dimensional full-horizon automatic tracking method | |
| CN102918423B (en) | Method for earthquake hydrocarbon system anlysis | |
| CN104914465B (en) | Volcanic Rock quantitative forecasting technique and device | |
| CN104597493B (en) | The evaluation methodology of observation system geometry changed design and device based on actual seismic data | |
| CN105528522A (en) | Method and device for calculating quantity of resources of continuous oil and gas reservoir based on space grids | |
| CN105093313B (en) | A kind of Karst-type oil reservoir individual well oil and gas productivity prediction method and device | |
| CN111596978A (en) | Web page display method, module and system for lithofacies classification by artificial intelligence | |
| Tolstukhin et al. | Ekofisk 4D seismic-seismic history matching workflow | |
| Li | Geophysical exploration technology: Applications in lithological and stratigraphic reservoirs | |
| CN106415321A (en) | Instantaneous isochron attribute-based geobody identification for reservoir modeling | |
| CN109763814B (en) | Stratum matching visual analysis method based on multi-dimensional logging data | |
| Ma et al. | Depositional facies analysis and modeling of the Judy Creek reef complex of the Upper Devonian Swan Hills, Alberta, Canada | |
| CN109633745A (en) | A kind of drafting method and device of three-dimensional structural map | |
| CN108663714A (en) | A kind of sedimentary micro depicting method | |
| Zhou et al. | Data driven modeling and prediction for reservoir characterization using seismic attribute analyses and big data analytics | |
| Korjani et al. | Reservoir characterization using fuzzy kriging and deep learning neural networks | |
| Li et al. | Seismic structure interpretation based on machine learning: A case study in coal mining | |
| Ismail et al. | Unsupervised machine learning and multi-seismic attributes for fault and fracture network interpretation in the Kerry Field, Taranaki Basin, New Zealand | |
| CN108508481A (en) | A kind of method, apparatus and system of longitudinal wave converted wave seismic data time match | |
| CN103543478A (en) | Geologic morphological interpolation KM (Kriging and Multiple-point geostatistics) method | |
| Maity et al. | Reservoir characterization of an unconventional reservoir by integrating microseismic, seismic, and well log data | |
| CN103278852B (en) | Utilize the method for seismic data volume waveform configuration characteristic model predicting oil/gas | |
| CN115880455A (en) | Three-dimensional intelligent interpolation method based on deep learning | |
| Lorentzen et al. | Mapping Cretaceous faults using a convolutional neural network-A field example from the Danish North Sea. | |
| Huang et al. | New seismic attribute: Fractal scaling exponent based on gray detrended fluctuation analysis |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20141210 |
|
| RJ01 | Rejection of invention patent application after publication |