为了充分挖掘叠前地震数据中反映地震相类别的细微信息,充分利用钻井、地质等先验知识,提出基于叠前地震纹理属性的半监督地震相分析算法.该算法首先引入叠前地震纹理属性,以突出地震反射信息中的微小空间、振幅随方位角/炮检距等的变异性;然后采用自组织映射,对训练样本对进行训练;最后在钻井先验知识的约束下,对自组织映射输出层的神经元进行半监督聚类,得到神经元与地震相类别的映射关系.理论模型和实际资料应用结果证实该方法可提高地震相图的准确度及地震微相的识别能力,是一种性能优越的地震相分析工具.

There are some problems in conventional seismic facies analysis methods, such as easily plunge into local optimal solution, low sensitivity and without using prior knowledge. To solve the above-mentioned problems, we propose a pre-stack texture-based semi-supervised seismic facies analysis method with global optimization. Firstly, the pre-stack seismic texture attributes are introduced to highlighting the information of micro-spatial and amplitude variation with azimuth/offset in seismic reflection data. Then, the self-organizing map (SOM) neural network is used to compress a large amount of redundant information of the samples on the premise of maintaining the topology of the data. Finally, the artificial bee colony (ABC) algorithm is used to realize the global optimization of the clustering of neurons in the SOM output layer under the constraints of prior knowledge. Besides, according to the probability estimation results based on the probabilistic neural network (PNN), we define the confidence measures to quantitative analysis the classification results. The synthetic test and practical application results show that the proposed method can not only significantly improve the recognition ability of the seismic microfacies, but also improve the horizontal resolution and the accuracy of the seismic facies map. These satisfactory results illustrate the proposed method is an effective tool for seismic facies analysis.

PreviousNext No AccessInternational Geophysical Conference, Beijing, China, 24-27 April 2018Semi-supervised fast algorithm for seismic waveform classificationAuthors: Hanpeng CaiHaiyang RenQingping WuLongkang PengChuanyong WenHanpeng CaiSchool of Resources and Environment, University of Electronic Science and Technology of China, and Center for Information Geoscience, University of Electronic Science and Technology of ChinaSearch for more papers by this author, Haiyang RenSchool of Resources and Environment, University of Electronic Science and Technology of ChinaSearch for more papers by this author, Qingping WuSchool of Resources and Environment, University of Electronic Science and Technology of ChinaSearch for more papers by this author, Longkang PengSchool of Resources and Environment, University of Electronic Science and Technology of ChinaSearch for more papers by this author, and Chuanyong WenSchool of Resources and Environment, University of Electronic Science and Technology of ChinaSearch for more papers by this authorhttps://doi.org/10.1190/IGC2018-287 SectionsAboutPDF/ePub ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InRedditEmail Abstract Waveform classification technique is a powerful tool for seismic facies analysis and reservoir prediction. The existing seismic waveform classification method is mainly based on unsupervised classification algorithm, without considering drilling, logging, geological and other priori knowledge. To make full use of these priori knowledge and improve the efficiency of computational classification, a semi-supervised fast seismic waveform classification algorithm is introduced in this paper. By using the semi-supervised dimensionality reduction (SSDR) algorithm based on liner transformation, we could accomplish three goals: reducing the dimension of the samples which need to be classified, enhancing the similarity of the samples in the same category and highlighting the characteristics of the different samples. On the basis of dimensionality reduction, the semi-supervised K-means algorithm based on a new distance measurement can obtain the seismic facies map. After testing the proposed method in the actual data, the results show that the proposed method can significantly improve the accuracy and efficiency of seismic waveform classification, and be an alternative better tool for semi-supervised seismic facies analysis. Keywords: algorithm, logging, faciesPermalink: https://doi.org/10.1190/IGC2018-287FiguresReferencesRelatedDetailsCited byAutomated Platform for Microseismic Signal Analysis: Denoising, Detection, and Classification in Slope Stability StudiesIEEE Transactions on Geoscience and Remote Sensing, Vol. 59, No. 9 International Geophysical Conference, Beijing, China, 24-27 April 2018ISSN (online):2159-6832Copyright: 2018 Pages: 1821 publication data© 2018 Published in electronic format with permission by the Society of Exploration Geophysicists and Chinese Geophysical SocietyPublisher:Society of Exploration Geophysicists HistoryPublished Online: 11 Dec 2018 CITATION INFORMATION Hanpeng Cai, Haiyang Ren, Qingping Wu, Longkang Peng, and Chuanyong Wen, (2018), "Semi-supervised fast algorithm for seismic waveform classification," SEG Global Meeting Abstracts : 1175-1179. https://doi.org/10.1190/IGC2018-287 Plain-Language Summary KeywordsalgorithmloggingfaciesPDF DownloadLoading ...

There are very strong lateral and vertical heterogeneities in karst cave reservoirs. Effective feature extraction from seismic data becomes more challenging for identifying karst cave reservoirs. In this paper, we present an improved deep learning model using the optimized convolutional neural network (OCNN) to identifying karst caves. As a supervised deep learning method, the proposed method learns nonlinear, discriminant, and invariant features from labeled data set, in which a large number of labeled samples are required. We suggest that a great deal of seismic numerical modeling results are added to the actual labeled samples data set from filed seismic surveys to ensure the adequacy of labeled samples. Seismic data usually contaminated with a variety of noise, so we suggest that using de-noising auto-encoder (DAE) enhance OCNN’s robustness prior to OCNN training. Experimental results demonstrate comprehensively the effectiveness of the proposed scheme, and show that it provides a much higher accuracy than the conventional methods, such as Root Mean Square amplitude (RMS amplitude) and Self-organizing Maps (SOM). Presentation Date: Wednesday, October 17, 2018 Start Time: 1:50:00 PM Location: Poster Station 2 Presentation Type: Poster