CN103488971B - Method for identifying geometrical morphology of organic reef storage layer - Google Patents

Abstract

The invention discloses a method for identifying the geometrical morphology of an organic reef storage layer. The method includes the steps of firstly, conducting preprocessing on original seismic profile data, namely, selecting a sliding window and the movement step size of the sliding window according to the size of a seismic profile; secondly, conducting feature extraction on the seismic profile data preprocessed in the first step through the texture attributive feature extraction method derived on the basis of a gray level co-occurrence matrix or through the feature extraction method based on Fourier spectrum feathers; thirdly, conducting organic reef storage layer clustering and inorganic reef storage layer clustering on features extracted in the second step through the K-means clustering method based on the distance; fourthly, attaching clustering labels in different colors to the clustered organic reef storage layer and a clustered inorganic reef storage layer and displaying the labels on the seismic profile. According to the method for identifying the geometrical morphology of the organic reef storage layer, the seismic profile data are processed through a signal processing method, the mode identification technology is introduced to identify the organic reef storage layer, the identification rate of the organic reef storage layer is effectively improved, and the method is particularly suitable for processing large-scale data.

Description

The geometric shape recognition methodss of reef reservoir

Technical field

The invention belongs to geologic image processing technology field is and in particular to a kind of geometric shape identification side of reef reservoir Method.

Background technology

In very long earth history, from Archean microorganism algal reef to modern Corallium Japonicum Kishinouye and Corallium Japonicum Kishinouye algal reef it may be said that Organic reef is the product of earliest vital action on the earth.Organic reef is commonly formed the construction of protuberantia, along with after primary and diagenesis The porosity and permeability that raw effect produces, is conducive to petroleum gas to assemble, and after being covered by other sedimentary covers It is commonly formed good oil-bearing structure, the research of therefore organic reef has very important significance.

The research method being directed to reef reservoir in terms of seismic prospecting in recent years emerges in an endless stream, and its research is concentrated mainly on Following aspect：（1）The numerical simulation of seismic reservoir response characteristic；（2）Ancient landform is analyzed, such as formation thickness, flattening etc.；（3） Seismic facies analysis based on Seismic waveform classification；（4）Seismic attributes analysis；（5）Poststack seismic inversion is analyzed；（6）Time frequency analysis Method；（7）Seismic wave attenuation by absorption is analyzed；（8）Fluid replacement analysiss based on Gassman equation, petrophysical model；（9） Prestack elastic parameter inversion；（10）Fluid identification factorial analyses etc..But the drilling success of reef reservoir is still very low, this is Because its geophysics prediction difficulty is very big：

（1）Due to can be affected by factors such as geology, deposition, weathers during being formed, life can be shaped The area of space of thing reef reservoir is very harsh；

（2）Reef reservoir spatial distribution very irregular；

（3）The depth that reef reservoir buries is big, and the resolution on the seismic data obtaining and signal to noise ratio are relatively low；

（4）The type of organic reef is many, and internal structure is complicated, and geophysical response characteristic changes greatly.

Therefore, the existing research laying particular emphasis on stratum geology aspect can not identify reef reservoir well.

Content of the invention

It is an object of the invention to overcoming the problems referred to above of the prior art, one kind is provided to be based on texture properties, Fourier A kind of method of the identification reef reservoir geometric shape of the principle such as spectrum sigtral response and chaos texture properties.

For solving above-mentioned technical problem, the present invention employs the following technical solutions：

A kind of geometric shape recognition methodss of reef reservoir it is characterised in that：Comprise the following steps：

S1：Pretreatment is carried out to original earthquake cross-sectional data, chooses sliding window and cunning including according to seismic profile size The moving step length of dynamic window；

S2：Feature extraction is carried out to seismic profile data pretreated in S1；

S3：The feature that S2 is extracted carries out reef reservoir and the cluster of abiotic reef reservoir；

S4：To the reef reservoir after cluster and abiotic reef reservoir patch cluster labels in different colors to be shown in ground On shake section.

Further, carry out feature using the texture properties feature extraction derived based on gray level co-occurrence matrixes in S2 to carry Take.

Further, as extraction feature in the feature extraction of the described texture properties based on gray level co-occurrence matrixes derivation For gray level be 16, element be 0 ° of gray level co-occurrence matrixes of joint probability density of the pixel that two spacing d are 1,45 °, The average of four stack features parameters of 90 °, 135 ° four direction derivation.

Further, described characteristic parameter includes：Angular second moment（ASM）, contrast, correlation, entropy, unfavourable balance away from and homogeneity Tolerance.

Further, in described S1, original earthquake cross-sectional data is carried out first extracting original earthquake cross-sectional data before pretreatment Chaos texture properties.

Further, in S2, feature extraction is carried out using the feature extraction based on Fourier spectrum feature.

Further, described comprised the following steps based on the feature extraction of Fourier spectrum feature：

S21：The Fourier power spectrum of seismic section image pretreated in S1 is divided into several equidistantly concentric squares Shape ring, several are labeled as M；If the size of image f (x, y) is L × W, image is placed in cartesian coordinate system, makes figure The coordinate of inconocenter is（L/2, W/2）, wherein, L represents the length of image, and W represents the width of image, then the direct computation of DFT of image Leaf transformation is：

$F(u,v)=\underset{x=0}{\overset{L-1}{\mathrm{\Σ}}}\underset{y=0}{\overset{W-1}{\mathrm{\Σ}}}f(x,y)\exp [-j2\mathrm{\π}(\frac{\mathrm{ux}}{L}+\frac{\mathrm{vy}}{W})]=R(u,v)+\mathrm{jI}(u,v)---\left(1\right)$

Wherein, u=0 ..., L-1, v=0 ..., W-1；X, y are the horizontal stroke of image, axis of ordinates；F (u, v) is frequency domain figure picture Spectrum, usual F (u, v) is the complex function of two real frequency variable u and v, and frequency u corresponds to x-axis, and frequency v corresponds to y Axle；R (u, v) is the real part of F (u, v), and I (u, v) is the imaginary part of F (u, v)；

The energy spectrum of point (u, v) is：

p(u,v)=R²(u,v)+I²(u,v) （2）

S22：Obtain the energy of each straight-flanked ring power spectrum；

The energy of each straight-flanked ring power spectrum is：

E_i=∑ p (u, v) (3)

Its summation scope be：L/2M*i≤|u-L/2|<L/2M* (i+1),

W/2M*i≤|v-W/2|<W/2M*(i+1),i=0,1,2,3...M-1；

S23：Each straight-flanked ring power spectral energies is asked to account for the percentage ratio of all straight-flanked ring power spectrum gross energies；

The computing formula of percentage ratio such as formula（4）Shown：

${\mathrm{fh}}_i=E_i/{\mathrm{\&Sigma;}}_{u=0}^L{\mathrm{\&Sigma;}}_{v=0}^{W}p(u,v)---\left(4\right)$

By tried to achieve percentage ratio fh_iCarry out feature extraction as extracting feature.

Further, the moving step length of described sliding window is the 2/3 of the sliding window length of side.

Further, step S3 carries out biology using based on the feature that the K-means clustering method of distance is extracted to S2 Reef reservoir and the cluster of abiotic reef reservoir.

Compared with prior art, the invention has the beneficial effects as follows：

First, the geometric shape recognition methodss of reef reservoir proposed by the present invention utilize the method for signal processing to process ground Shake data, and introduce mode identification technology identification reef reservoir, effectively increase the discrimination of reef reservoir；

Secondly, the geometric shape recognition methodss of reef reservoir proposed by the present invention are by the region geometry shape of reef reservoir State identifies, provides technical support for studying developable reef reservoir further.

Finally, the geometric shape recognition methodss computation complexity of reef reservoir proposed by the present invention is low, it is possible to reduce big The complicated positioning work of amount, is particularly well-suited to the process of large-scale data.

Brief description

Fig. 1 is the flow chart of the geometric shape recognition methodss of reef reservoir in embodiment 1；

Fig. 2 is the flow chart of the geometric shape recognition methodss of reef reservoir in embodiment 2；

Fig. 3 is the flow chart of the geometric shape recognition methodss of reef reservoir in embodiment 3；

Fig. 4 is the original earthquake cross-sectional data in embodiment 3；

Fig. 5 is the chaos texture properties of the original earthquake cross-sectional data extracted in embodiment 3;

Fig. 6 is the seismic cross-section showing reef reservoir and abiotic reef reservoir label in embodiment 1,2,3.

Specific embodiment

In order that the objects, technical solutions and advantages of the present invention become more apparent, below in conjunction with drawings and Examples, right The present invention is further elaborated.It should be appreciated that specific embodiment described herein is only in order to explain the present invention, not For limiting the present invention.

Embodiment 1

As shown in the flowchart of fig.1, the geometric shape recognition methodss of the reef reservoir in the present embodiment include following step Suddenly：

S1：Pretreatment is carried out to original earthquake cross-sectional data, chooses sliding window and cunning including according to seismic profile size The moving step length of dynamic window, wherein, the moving step length of sliding window is the 2/3 of the sliding window length of side；

S2：Using the texture properties feature extraction derived based on gray level co-occurrence matrixes, earthquake pretreated in S1 is cutd open Face data carries out feature extraction；

In features described above extraction process, as extract feature for gray level be 16, element be the picture that two spacing d are 1 0 ° of the gray level co-occurrence matrixes of joint probability density of element, 45 °, 90 °, the four stack features parameters that derive of 135 ° of four directions equal Value, described characteristic parameter includes：Angular second moment（ASM）, contrast, correlation, entropy, unfavourable balance measure away from homogeneity.

It is prior art based on the texture properties feature extraction that gray level co-occurrence matrixes are derived, here does not do expansion explanation.

S3：Carry out reef reservoir and Fei Sheng using based on the feature that the K-means clustering method of distance is extracted to S2 The cluster of thing reef reservoir；

S4：To the reef reservoir after cluster and abiotic reef reservoir patch cluster labels in different colors to be shown in ground On shake section, as shown in fig. 6, wherein, black portions represent reef reservoir region, and grey parts represent abiotic reef reservoir region Domain.

Embodiment 2

As shown in the flowchart of figure 2, the geometric shape recognition methodss of the reef reservoir in the present embodiment include following step Suddenly：

S1：Pretreatment is carried out to original earthquake cross-sectional data, chooses sliding window and cunning including according to seismic profile size The moving step length of dynamic window, wherein, the moving step length of sliding window is the 2/3 of the sliding window length of side；

S2：Using the feature extraction based on Fourier spectrum feature, seismic profile data pretreated in S1 is carried out Feature extraction, specifically includes following steps：

S21：The Fourier power spectrum of seismic section image pretreated in S1 is divided into several equidistantly concentric squares Shape ring, several are labeled as M；If the size of image f (x, y) is L × W, image is placed in cartesian coordinate system, makes figure The coordinate of inconocenter is（L/2, W/2）, wherein, L represents the length of image, and W represents the width of image, then the direct computation of DFT of image Leaf transformation is：

$F(u,v)=\underset{x=0}{\overset{L-1}{\mathrm{\&Sigma;}}}\underset{y=0}{\overset{W-1}{\mathrm{\&Sigma;}}}f(x,y)\exp [-j2\mathrm{\&pi;}(\frac{\mathrm{ux}}{L}+\frac{\mathrm{vy}}{W})]=R(u,v)+\mathrm{jI}(u,v)---\left(1\right)$

Wherein, u=0 ..., L-1, v=0 ..., W-1；X, y are the horizontal stroke of image, axis of ordinates；F (u, v) is frequency domain figure picture Spectrum, usual F (u, v) is the complex function of two real frequency variable u and v, and frequency u corresponds to x-axis, and frequency v corresponds to y Axle；R (u, v) is the real part of F (u, v), and I (u, v) is the imaginary part of F (u, v).

The energy spectrum of point (u, v) is：

p(u,v)=R²(u,v)+I²(u,v) （2）

S22：Obtain the energy of each straight-flanked ring power spectrum：

The energy of each straight-flanked ring power spectrum is：

E_i=∑p(u,v) (3)

Its summation scope be：L/2M*i≤|u-L/2|<L/2M*(i+1),

W/2M*i≤|v-W/2|<W/2M*(i+1),i=0,1,2,3...M-1；

S23：Each straight-flanked ring power spectral energies is asked to account for the percentage ratio of all straight-flanked ring power spectrum gross energies；

The computing formula of percentage ratio is as follows：

${\mathrm{fh}}_i=E_i/{\mathrm{\&Sigma;}}_{u=0}^L{\mathrm{\&Sigma;}}_{v=0}^{W}p(u,v)---\left(4\right)$

S24：Gained straight-flanked ring power spectral energies are accounted for the percentage ratio fh of all straight-flanked ring power spectrum gross energies_iAs extraction Feature carries out feature extraction.

S3：Carry out reef reservoir and Fei Sheng using based on the feature that the K-means clustering method of distance is extracted to S2 The cluster of thing reef reservoir；

S4：To the reef reservoir after cluster and abiotic reef reservoir patch cluster labels in different colors to be shown in ground On shake section, as shown in fig. 6, wherein, black portions represent reef reservoir region, and grey parts represent abiotic reef reservoir region Domain.

Embodiment 3

As depicted in the flow chart of fig.3, the geometric shape recognition methodss of the reef reservoir in the present embodiment include following step Suddenly：

S1：Original earthquake cross-sectional data is carried out with chaos texture properties extraction, extracts result as shown in figure 5, Fig. 4 show Original earthquake cross-sectional data；

S2：Pretreatment is carried out to S1 the data obtained, chooses sliding window and sliding window including according to seismic profile size Moving step length, wherein, the moving step length of sliding window is the 2/3 of the sliding window length of side；

S3：Using the texture properties feature extraction derived based on gray level co-occurrence matrixes, earthquake pretreated in S1 is cutd open Face data carries out feature extraction；

In features described above extraction process, as extract feature for gray level be 16, element be the picture that two spacing d are 1 0 ° of the gray level co-occurrence matrixes of joint probability density of element, 45 °, 90 °, the four stack features parameters that derive of 135 ° of four directions equal Value, characteristic parameter includes：Angular second moment（ASM）, contrast, correlation, entropy, unfavourable balance measure away from homogeneity.

It is prior art based on the texture properties feature extraction that gray level co-occurrence matrixes are derived, here does not do expansion explanation.

S4：Carry out reef reservoir and Fei Sheng using based on the feature that the K-means clustering method of distance is extracted to S3 The cluster of thing reef reservoir；

S5：To the reef reservoir after cluster and abiotic reef reservoir patch cluster labels in different colors to be shown in ground On shake section, as shown in fig. 6, wherein, black portions represent reef reservoir region, and grey parts represent abiotic reef reservoir region Domain.

Those of ordinary skill in the art will be appreciated that, embodiment described here is to aid in reader and understands this Bright principle is it should be understood that protection scope of the present invention is not limited to such special statement and embodiment.This area Those of ordinary skill can make various other each without departing from present invention essence according to these technology disclosed by the invention enlightenment Plant concrete deformation and combine, these deform and combine still within the scope of the present invention.

Claims (5)

1. a kind of reef reservoir geometric shape recognition methodss it is characterised in that：Comprise the following steps：

S1：Pretreatment is carried out to original earthquake cross-sectional data, chooses sliding window and sliding window including according to seismic profile size The moving step length of mouth；

S2：Feature extraction is carried out to seismic profile data pretreated in S1,

Specifically feature extraction is carried out using the texture properties feature extraction derived based on gray level co-occurrence matrixes；

Or,

Specifically feature extraction is carried out using the feature extraction based on Fourier spectrum feature；

In the feature extraction of the described texture properties based on gray level co-occurrence matrixes derivation as extract feature for gray level it is 16th, element is 0 °, 45 °, 90 °, 135 ° four of the gray level co-occurrence matrixes of the joint probability density of pixel that two spacing d are 1 The average of the four stack features parameters that direction is derived；

Described comprised the following steps based on the feature extraction of Fourier spectrum feature：

S21：The Fourier power spectrum of seismic section image pretreated in S1 is divided into several equidistant concentric rectangles rings, Several are labeled as M；If the size of image f (x, y) is L × W, image is placed in cartesian coordinate system, makes in image The coordinate of the heart is (L/2, W/2), and wherein, L represents the length of image, and W represents the width of image, then the discrete fourier of image becomes It is changed to：

$F(u,v)=\underset{x=0}{\overset{L-1}{\&Sigma;}}\underset{y=0}{\overset{W-1}{\&Sigma;}}f(x,y)\exp \&lsqb;-j2\mathrm{\&pi;}(\frac{ux}{L}+\frac{vy}{W})\&rsqb;=R(u,v)+jI(u,v)---\left(1\right)$

Wherein, u=0 ..., L-1, v=0 ..., W-1；X, y are the horizontal stroke of image, axis of ordinates；F (u, v) is frequency domain figure picture spectrum, Generally F (u, v) is the complex function of two real frequency variable u and v, and frequency u corresponds to x-axis, and frequency v corresponds to y-axis；R (u, v) is the real part of F (u, v), and I (u, v) is the imaginary part of F (u, v)；

The energy spectrum of point (u, v) is：

P (u, v)=R²(u,v)+I²(u,v) (2)

S22：Obtain the energy of each straight-flanked ring power spectrum；

The energy of each straight-flanked ring power spectrum is：

E_i=∑ p (u, v) (3)

Its summation scope be：L/2M*i≤| u-L/2 | ＜ L/2M* (i+1),

W/2M*i≤| v-W/2 | ＜ W/2M* (i+1), i=0,1,2,3 ..., M-1；

S23：Each straight-flanked ring power spectral energies is asked to account for the percentage ratio of all straight-flanked ring power spectrum gross energies；

Shown in the computing formula of percentage ratio such as formula (4)：

${\mathrm{fh}}_i=E_i/{\mathrm{\&Sigma;}}_{u=0}^L{\mathrm{\&Sigma;}}_{v=0}^{W}p(u,v)---\left(4\right)$

By tried to achieve percentage ratio fh_iCarry out feature extraction as extracting feature；

S3：The feature that S2 is extracted carries out reef reservoir and the cluster of abiotic reef reservoir；

S4：Reef reservoir after cluster and abiotic reef reservoir patch cluster labels in different colors are cutd open with being shown in earthquake On face.

2. reef reservoir according to claim 1 geometric shape recognition methodss it is characterised in that：Described characteristic parameter Including：Angular second moment (ASM), contrast, correlation, entropy, unfavourable balance are measured away from homogeneity.

3. reef reservoir according to claim 1 geometric shape recognition methodss it is characterised in that：To former in described S1 Beginning seismic profile data first extracts the chaos texture properties of original earthquake cross-sectional data before carrying out pretreatment.

4. reef reservoir according to claim 1 geometric shape recognition methodss it is characterised in that：Described sliding window Moving step length be the sliding window length of side 2/3.

5. reef reservoir according to claim 1 geometric shape recognition methodss it is characterised in that：Step S3 adopts base Carry out reef reservoir and the cluster of abiotic reef reservoir in the feature that the K-means clustering method of distance is extracted to S2.