多属性数据中基于连续平行坐标的可视分析方法研究
发布时间：2018-11-19 19:02
【摘要】：多维多属性数据分析和处理是海量数据分析的重要内容之一。在油气资源勘探领域,由于信号的信噪比较低,单个属性中的目标特征不明显,通过多属性融合分析,可以凸显地质构造特征和地质目标特征。基于此,本文针对多属性地震数据问题开展研究,提出了基于可视分析的多属性地震数据分析方法,其基本思想是结合可视化技术和人机交互技术,充分利用计算机的处理能力和人的主观经验,一方面可以避免单纯依赖计算机进行数据分析的准确性问题,另一方面可以避免仅依赖人机交互带来的操作复杂性问题。具有一定的理论价值和实际应用价值。本文针对多属性地震数据的可视分析问题开展研究,主要贡献如下:1.提出了基于连续平行坐标的多属性数据可视分析方法。针对多属性数据的可视分析,提出了多属性数据可视化、人机交互的特征提取和融合体绘制的可视分析流程。首要问题是对大规模多属性数据的可视化问题,本文采用了基于连续平行坐标的多维多属性可视化方法实现了多属性数据的展示,提取和凸显目标的特征。通过人机交互流程实现目标特征的迭代分析和提取,在此基础上将目标特征映射成融合体绘制的传递函数,在三维空间采用融合体绘制技术将目标特征进行展示。人机交互贯穿整个数据分析流程,实现了实时的多属性可视分析方法。通过仿真分析,本文提出的方法可有有效解决多多属性地震数据的分析问题;2.提出了基于空间信息的多属性数据可视分析方法。多属性数据中,不同属性具有一定的相关性,同时,目标特征在空间上具有一定的相关性和连续性。基于此,本文提出了基于空间信息的多属性数据可视分析方法。其基本思想是通过人机交互拾取地质目标的局部信息,利用目标在空间上的相关性和连续性,凸显目标的空间特征。基本过程是将多属性值在散点图中投影,用散点图中投影到每个像素的体素的重心坐标和空间方差来表征体素的空间信息,并根据空间信息来对数据进行分类,从而设计传递函数指导融合体绘制结果。该方法在保留地质目标特征的同时,消除了非特征物质的干扰,提高了特征提取效果;3.设计并实现了集成这两种方法的可视分析系统。利用实际的地震多属性数据进行仿真,为本文的方法提供实际验证。本文针对海量多属性数据的分析问题开展研究,并提出可视分析方法。通过仿真分析,本文提出的方法有效解决多属性地震数据的可视分析问题。
[Abstract]:Multi-dimensional and multi-attribute data analysis and processing is one of the important contents of mass data analysis. In the field of oil and gas resource exploration, because of the low signal-to-noise ratio of the signal, the target characteristics in single attribute are not obvious. Through the multi-attribute fusion analysis, the geological structure characteristics and geological target characteristics can be highlighted. Based on this, this paper studies the problem of multi-attribute seismic data, and puts forward a method of multi-attribute seismic data analysis based on visual analysis. The basic idea of this method is to combine visualization technology and human-computer interaction technology. Making full use of the computer's processing power and human's subjective experience, on the one hand, we can avoid the accuracy problem of data analysis only relying on the computer, on the other hand, we can avoid the operational complexity problem caused by only relying on human-computer interaction. It has certain theoretical value and practical application value. The main contributions of this paper are as follows: 1. A method for visual analysis of multi-attribute data based on continuous parallel coordinates is proposed. Aiming at visual analysis of multi-attribute data, a visual analysis flow of multi-attribute data visualization, feature extraction of human-computer interaction and fusion volume rendering is proposed. The most important problem is the visualization of large-scale multi-attribute data. In this paper, the multi-attribute visualization method based on continuous parallel coordinates is used to display the multi-attribute data, extract and highlight the features of the target. Based on the iterative analysis and extraction of target features through human-computer interaction process, the target features are mapped to transfer functions of fusion volume rendering, and the target features are displayed by fusion volume rendering technology in three-dimensional space. The man-machine interaction runs through the whole data analysis flow and realizes the real-time multi-attribute visual analysis method. Through simulation analysis, the method proposed in this paper can effectively solve the problem of seismic data analysis with many attributes. 2. A method for visual analysis of multi-attribute data based on spatial information is proposed. In multi-attribute data, different attributes have a certain correlation, at the same time, the target features have a certain correlation and continuity in space. Based on this, this paper presents a visual analysis method of multi-attribute data based on spatial information. The basic idea is to pick up the local information of the geological target through human-computer interaction, and to highlight the spatial characteristics of the target by using the spatial correlation and continuity of the target. The basic process is to project the multi-attribute value in the scatter plot, and use the barycenter coordinate and spatial variance of the voxel projected to each pixel in the scattered plot to represent the spatial information of the voxel, and classify the data according to the spatial information. Thus the transfer function is designed to guide the fusion volume rendering results. This method not only preserves the geological target features, but also eliminates the interference of non-characteristic substances and improves the feature extraction effect. 3. A visual analysis system integrating these two methods is designed and implemented. The simulation is carried out by using the actual seismic multi-attribute data, which provides the practical verification for the method in this paper. In this paper, the analysis of massive multi-attribute data is studied, and a visual analysis method is proposed. Through simulation analysis, the method proposed in this paper can effectively solve the problem of visual analysis of multi-attribute seismic data.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P618.13;P631.44
本文编号：2343163
[Abstract]:Multi-dimensional and multi-attribute data analysis and processing is one of the important contents of mass data analysis. In the field of oil and gas resource exploration, because of the low signal-to-noise ratio of the signal, the target characteristics in single attribute are not obvious. Through the multi-attribute fusion analysis, the geological structure characteristics and geological target characteristics can be highlighted. Based on this, this paper studies the problem of multi-attribute seismic data, and puts forward a method of multi-attribute seismic data analysis based on visual analysis. The basic idea of this method is to combine visualization technology and human-computer interaction technology. Making full use of the computer's processing power and human's subjective experience, on the one hand, we can avoid the accuracy problem of data analysis only relying on the computer, on the other hand, we can avoid the operational complexity problem caused by only relying on human-computer interaction. It has certain theoretical value and practical application value. The main contributions of this paper are as follows: 1. A method for visual analysis of multi-attribute data based on continuous parallel coordinates is proposed. Aiming at visual analysis of multi-attribute data, a visual analysis flow of multi-attribute data visualization, feature extraction of human-computer interaction and fusion volume rendering is proposed. The most important problem is the visualization of large-scale multi-attribute data. In this paper, the multi-attribute visualization method based on continuous parallel coordinates is used to display the multi-attribute data, extract and highlight the features of the target. Based on the iterative analysis and extraction of target features through human-computer interaction process, the target features are mapped to transfer functions of fusion volume rendering, and the target features are displayed by fusion volume rendering technology in three-dimensional space. The man-machine interaction runs through the whole data analysis flow and realizes the real-time multi-attribute visual analysis method. Through simulation analysis, the method proposed in this paper can effectively solve the problem of seismic data analysis with many attributes. 2. A method for visual analysis of multi-attribute data based on spatial information is proposed. In multi-attribute data, different attributes have a certain correlation, at the same time, the target features have a certain correlation and continuity in space. Based on this, this paper presents a visual analysis method of multi-attribute data based on spatial information. The basic idea is to pick up the local information of the geological target through human-computer interaction, and to highlight the spatial characteristics of the target by using the spatial correlation and continuity of the target. The basic process is to project the multi-attribute value in the scatter plot, and use the barycenter coordinate and spatial variance of the voxel projected to each pixel in the scattered plot to represent the spatial information of the voxel, and classify the data according to the spatial information. Thus the transfer function is designed to guide the fusion volume rendering results. This method not only preserves the geological target features, but also eliminates the interference of non-characteristic substances and improves the feature extraction effect. 3. A visual analysis system integrating these two methods is designed and implemented. The simulation is carried out by using the actual seismic multi-attribute data, which provides the practical verification for the method in this paper. In this paper, the analysis of massive multi-attribute data is studied, and a visual analysis method is proposed. Through simulation analysis, the method proposed in this paper can effectively solve the problem of visual analysis of multi-attribute seismic data.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P618.13;P631.44
【参考文献】
相关期刊论文 前4条
1 凌云;高军;孙德胜;林吉祥;;裂缝储层的地震预测技术与应用实例[J];石油地球物理勘探;2015年01期
2 王开燕;徐清彦;张桂芳;程某存;李培海;;地震属性分析技术综述[J];地球物理学进展;2013年02期
3 毕达天;邱长波;张晗;;数据降维技术研究现状及其进展[J];情报理论与实践;2013年02期
4 张延玲;杨长春;贾曙光;;地震属性技术的研究和应用[J];地球物理学进展;2005年04期
相关硕士学位论文 前1条
1 翟旭君;基于平行坐标的可视化数据挖掘技术研究[D];清华大学;2005年
，本文编号：2343163
本文链接：https://www.wllwen.com/kejilunwen/diqiudizhi/2343163.html
