作　　者： ; ;

机构地区： 广州大学数学与信息科学学院

出　　处： 《地球科学进展》 2007年第4期357-361,共5页

摘　　要： 自仿射分形的Hurst指数是分析地质剖面数据的有利参数。以大尹格庄金矿不同勘探线刻槽取样所得的金品位序列为例,评价Hurst指数的几种估算方法在地质剖面数据分析中的适用性。取相同尺度,全部数据集的增量标准偏差法统计散点呈波状变化,部分数据集的曲线长度变换法统计散点的后半部分呈波状变化,全部数据集的重标极差分析法统计散点线性拟合较好。结果显示增量标准偏差法对尺度要求较为苛刻,适于巨量数据的统计;曲线长度变换法应用性较广,所得Hurst指数反映品位的空间变化强度;重标极差分析法稳定性最好,其Hurst指数反映了品位变化相依性。 Hurst parameter of self-affine fractal is the favorable parameters for studying the data of geological profiles. Hurst parameters of data series of gold grade, derived from notch grooves in different exploratory lines, are calculated by the methods of incremental standard deviation, curve length transform and rescaled range analysis（ R/ S） , and the applicability of these methods is evaluated affectively. As uniform scaling analysis taken in the three methods, the plots by the first method show poor linear relationship in all the applications, the rear points obtained from the second method fluctuate in part of the applications, and the plots by the last method show good linear relationship in the most applications. Therefore, the restrictions of the dataset capacitance and the analysis scale for the incremental standard deviation method are more than those for the curve length transform method, and the R/S method can be applied widely than the two other methods and it is suitable for statistic analysis of small dataset. Moreover, the Hurst parameter received from the curve length transform method can reflect the spatial fluctuation intensity of gold grade, while the same parameter obtained by R/S analysis judges the persistence of grade distribution.

关 键 词： 自仿射分形 指数 相依性 成矿元素 胶东

领　　域： [天文地球] [天文地球]