基于多特征模糊聚类的图像融合方法

首先利用模糊C-均值聚类算法在多特征形成的特征空间上对图像进行区域分割,并在此基础上对区域进行多尺度小波分解;然后利用柯西函数构造区域的模糊相似度,应用模糊相似度及区域信息量构造加权因子,从而得到融合图像的小波系数;最后利用小波逆变换得到融合图像·采用均方根误差、峰值信噪比、熵、交叉熵和互信息5种准则评价融合算法的性能·实验结果表明,文中方法具有良好的融合特性·

地震岩性与物性联合反演技术研究和应用

This paper studies how to more effectively invert seismic data and predict reservoir under complicated sedimentary environment, complex rock physical relationships and fewer drills in offshore areas of China. Based on rock physical and seismic amplitude-preserving process, and according to depositional system and laws of hydrocarbon reservoir, in the light of feature of seismic inversion methods present applied, series methods were studied. A joint inversion technology for complex geological condition had been presented, at the same time the process and method system for reservoir prediction had been established. This method consists four key parts. 1)We presented the new conception called generalized wave impedance, established corresponding inversion process, and provided technical means for joint inversion lithology and petrophysical on complex geological condition. 2)At the aspect of high-resolution nonlinear seismic wave impedance joint inversion, this method used a multistage nonlinear seismic convolution model rather than conventional primary structure Robinson seismic convolution model, and used Caianiello neural network implement inversion. Based on the definition of multistage positive and negative wavelet, it adopted both deterministic and statistical physical mechanism, direct inversion and indirect inversion. It integrated geological knowledge, rock physical theory, well data, and seismic data, and improved the resolution and anti-noise ability of wave impedence inversion. 3)At the aspect of high-resolution nonlinear reservoir physical property joint inversion, this method used nonlinear rock physical model which introduced convolution model into the relationship between wave impedance and porosity/clay. Through multistage decomposition, it handles separately the large- and small-scale components of the impedance-porosity/clay relationships to achieve more accurate rock physical relationships. By means of bidirectional edge detection with wavelets, it uses the Caianiello neural network to finish statistical inversion with combined applications of model-based and deconvolution-based methods. The resulted joint inversion scheme can integrate seismic data, well data, rock physical theory, and geological knowledge for estimation of high-resolution petrophysical parameters. 4)At the aspect of risk assessment of lateral reservoir prediction, this method integrated the seismic lithology identification, petrophysical prediction, multi-scale decomposition of petrophysical parameters, P- and H-spectra, and the match relationship of data got from seismics, well logging and geology. It could describe the complexity of medium preferably. Through applications of the joint inversion of seismic data for lithologic and petrophysical parameters in several selected target areas, the resulted high-resolution lithologic and petrophysical sections(impedance, porosity, clay) show that the joint inversion can significantly improve the spatial description of reservoirs in data sets involving complex deposits. It proved the validity and practicality of this method adequately.

基于独立分量分析的地震属性分析方法研究

Conventional seismic attribute analysis is not only time consuming, but also has several possible results. Therefore, seismic attribute optimization and multi-attribute analysis are needed. In this paper, Fuyu oil layer in Daqing oil field is our main studying object. And there is much difference between seismic attributes and well logs. So under this condition, Independent Component Analysis (ICA) and Kohonen neural net are introduced to seismic attribute optimization and multi-attribute analysis. The main contents are as follows: (1) Now the method of seismic attribute compression is mainly principal component analysis (PCA). In this article, independent component analysis (ICA), which is superficially related to PCA, but much more powerful, is used to seismic reservoir characterizeation. The fundamental, algorithms and applications of ICA are surveyed. And comparation of ICA with PCA is stydied. On basis of the ne-entropy measurement of independence, the FastICA algorithm is implemented. (2) Two parts of ICA application are included in this article: First, ICA is used directly to identify sedimentary characters. Combined with geology and well data, ICA results can be used to predict sedimentary characters. Second, ICA treats many attributes as multi-dimension random vectors. Through ICA transform, a few good new attributes can be got from a lot of seismic attributes. Attributes got from ICA optimization are independent. (3) In this paper, Kohonen self-organizing neural network is studied. First, the characteristics of neural network’s structure and algorithm is analyzed in detail, and the traditional algorithm is achieved which has been used in seism. From experimental results, we know that the Kohonen self-organizing neural network converges fast and classifies accurately. Second, the self-organizing feature map algorithm needs to be improved because the result of classification is not very exact, the boundary is not quite clear and the velocity is not fast enough, and so on. Here frequency sensitive principle is introduced. Combine it with the self-organizing feature map algorithm, then get frequency sensitive self-organizing feature map algorithm. Experimental results show that it is really better. (4) Kohonen self-organizing neural network is used to classify seismic attributes. And it can be avoided drawing confusing conclusions because the algorithm’s characteristics integrate many kinds of seismic features. The result can be used in the division of sand group’s seismic faces, and so on. And when attributes are extracted from seismic data, some useful information is lost because of difference and deriveative. But multiattributes can make this lost information compensated in a certain degree.

整合多尺度地球物理数据建立高分辨率储层模型的方法及应用

Most of the fields in China are in the middle-late development phase or are mature fields. It becomes more and more difficult to develop the remaining oil/gas. Therefore, it is import to enhance oil/gas recovery in order to maintain the production. Fine scale modeling is a key to improve the recovery. Incorporation of geological, seismic and well log data to 3D earth modeling is essential to build such models. In Ken71 field, well log, cross-well seismic and 3D seismic data are available. A key issue is to build 3D earth model with these multi-scales data for oil field development.In this dissertation, studies on sequential Gaussian-Bayesian simulation have been conducted. Its comparison with cokriging and sequential Gaussian simulation has been performed. The realizations generated by sequential Gaussian-Bayesian simulation have higher vertical resolution than those generated by other methods. Less differences between these realization and true case are observed. With field data, it is proved that incorporating well log, cross-well seismic and 3D seismic into 3D fine scale model is reliable. In addition, the advantages of sequential Gaussian-Bayesian simulation and conditions for input data are demonstrated. In Ken71 field, the impedance difference between sandstone and shale is small. It would be difficult to identify sandstone in the reservoir with traditional impedance inversion. After comparisons of different inversion techniques, stochastic hillclimbing inversion was applied. With this method, shale content inversion is performed using 3D seismic data. Then, the inverted results of shale content and well log data are incorporated into 3D models. This demonstrates a procedure to build fine scale models using multi scale seismic data, especially 3D seismic amplitude volume.The models generated through sequential Gaussian-Bayesian simulation have several advantages including: (1) higher vertical resolution compared with 3D inverted acoustic impedance (AI); (2) consistency of lateral variation as 3D inverted AI; (3) more reliability due to integration cross-well seismic data. It is observed that the precision of the model depends on the 3D inversion.

松辽盆地古龙地区葡萄花油层组沉积成岩相研究

The sediment and diagenesis process of reservoir are the key controlling factors for the formation and distribution of hydrocarbon reservoir. For quite a long time, most of the research on sediment-diagenesis facise is mainly focusing on qualitative analysis. With the further development on exploration of oil field, the qualitative analysis alone can’t meet the requirements of complicated requirements of oil and gas exploreation, so the quantitative analysis of sediment-diagenesis facise and related facies modling have become more and more important. On the basis of the research result from stratum and sediment on GuLong Area Putaohua Oil Layer Group, from the basic principles of sedimentology, and with the support from the research result from field core and mining research results, the thesis mainly makes the research on the sediment types, the space framework of sands and the evolution rules of diagenesis while mainly sticking to the research on sediment systement analysis and diagenetic deformation, and make further quantitative classification on sediment-diageneses facies qualitatively, discussed the new way to divide the sediment-diagenesis facies, and offer new basis for reservoir exploration by the research. Through using statistics theory including factor analysis, cluster analysis and discriminant analysis, the thesis devided sediment-diagenesis facies quantitatively. This research method is innovative on studying sediment-diagenesis facies. Firstly, the factor analysis could study the main mechanism of those correlative variables in geologic body, and then could draw a conclusion on the control factors of fluid and capability of reservoir in the layer of studying area. Secondly, with the selected main parameter for the cluster analysis, the classification of diagenesis is mainly based on the data analysis, thus the subjective judgement from the investigator could be eliminated, besides the results could be more quantitative, which is helpful to the correlative statistical analysis, so one could get further study on the quantitative relations of each sediment-diagenesis facies type. Finally, with the reliablities of discriminant analysis cluster results, and the adoption of discriminant probability to formulate the chart, the thesis could reflect chorisogram of sediment-diagenesis facies for planar analysis, which leads to a more dependable analytic results.According to the research, with the multi-statistics analysis methods combinations, we could get quantitative analysis on sediment-diagenesis facies of reservoir, and the final result could be more reliable and also have better operability.

尺度匹配方法在井震联合重建和多次波衰减中的应用

Scale matching method means adjusting information with different scale to the same level. This thesis focuses on scale unification of information with different frequency bandwidth. Well-seismic cooperate inversion is an important component of reservoir geophysics; multiple prediction & subtraction is a development of multiple attenuation in recent years. The common ground of these two methods is that they both related to different frequency bandwidth unification. Well log、cross-hole seismic、VSP、3D seismic and geological information have different spatial resolution, we can decrease multi-solution of reservoir inversion and enhance the vertical and lateral resolution of the geological object by integrate those information together; Compare the predicted multiple generated by SRME with the real multiple, we find the predicted multiple convolutes at least one wavelet more, which brings frequency bandwidth difference between them. So the subtraction method also relates to multi-scale information unification. This thesis gives a method of well constrained seismic high resolution processing basing on auto gain control modulation. It uses base function method which utilizes original well-seismic match result as initial condition and processed seismic trace as initial model to extrapolate the high frequency information of the well logs to the seismic profiles. In this way we can broaden the bandwidth of the seismic and make the high frequency gain geological meaning. In this thesis we introduce the revised base function method to adaptive subtraction and verify the validity of the method using models. Key words: high frequency reconstruction, scale matching, base function, multiple, SRME prediction & subtraction

叠前地震多角度三参数同步反演方法

Pre-stack seismic inversion has become the emphasis and hotspot owing to the exploration & exploitation of oil field and the development of seismic technology. Pre-stack seismic inversion has the strongpoint of making the most of amplitude versus offset compared with the post-stack method. In this dissertation, the three parameters were discussed from multi-angle reflectance of P-wave data based on Zoeppritz’s and Aki & Richard’s equation, include P-wave velocity, S-wave velocity, and density. The three parameters are inversed synchronously from the pre-stack multi-angle P-wave data, based on rockphysics model and aimed at the least remnant difference between model simulation and practical data. In order to improve the stability of inversion and resolution to thin bed, several techniques were employed, such as the wavelet transform with multi-scale function, adding the Bayesian soft constraint and hard constraints (the horizon, structure and so on) to the inversion process. Being the result, the uncertainty of the resolution is reduced, the reliability and precision are improved, the significance of parameters becomes clearer. Meeting to the fundamental requirement of pre-stack inversion, some research in rockphysics are carried out which covered the simulation and inversion of S-wave velocity, the influence of pore fluids to geophysical parameters, and the slecting and analyzing of sensitive parameters. The difference between elastic wave equation modeling and Zoeppritz equation method is also compared. A series of key techniques of pre-stack seismic inversion and description were developed, such as attributes optimization, fluid factors, etc. All the techniques mentioned above are assembled to form a technique sets and process of synchronous pre-stack seismic inversion method of the three parameters based on rock physics and model simulation. The new method and technology were applied in many areas with various reservoirs, obtained both geological and economic significance, which proved to be valid and rational. This study will promote the pre-stack inversion technology and it’s application in hidden reservoirs exploration, face good prospects for development and application.