一种基于岩石薄片图像的粒度分析新方法  

A new method of determining grain size based on rock section image

作　　者：袁瑞[1,2] 朱锐[1] 瞿建华[3] 孙玉秋[2] 唐勇[3] 潘进[1] 

机构地区：[1]长江大学地球科学学院,武汉430100 [2]长江大学信息与数学学院,湖北荆州434023 [3]中国石油新疆油田分公司勘探开发研究院,新疆克拉玛依834000

出　　处：《岩性油气藏》2015年第5期104-107,共4页Lithologic Reservoirs

基　　金：国家自然科学基金"珠江口盆地渐新世-中新世潮汐沉积与潮汐周期研究"(编号:41302096);湖北省自然科学基金"随钻跟踪海量数据处理与成图关键技术研究"(编号:2013CFA053)联合资助

摘　　要：鉴于岩石薄片粒度分析的实验室技术具有较大的局限性,且速度缓慢,提出了一种基于岩石薄片图像空间自相关系数的粒度分析新方法。该方法通过计算已知颗粒大小的岩石薄片的空间自相关系数,并利用带约束条件的最小二乘法求解对应的空间自相关系数方程组,从而得到颗粒大小未知的岩石薄片的粒度分布。采用椭圆随机生成的方法制作了颗粒大小不同的理论模拟岩石薄片,并分析了其空间自相关系数与粒度的关系,得出空间自相关系数随着偏移距离的增大而减小;当偏移距离一定时,空间自相关系数随着粒度的增大而增大。采用该方法对粗砂岩中粒径分别为0.5~1.0 mm,1.0~1.5 mm和1.5~2.0 mm的颗粒进行了粒度分布的定量计算,得出这3种粒径颗粒所占的百分比分别为55.8%,24.6%和20.2%,该结果与实际值较接近,变化趋势与实际一致。Despite technological advances in lab instruments, grain-size analysis has many limitations, such as low speed.A theoretical method of determining grain size based on spatial autocorrelation coefficient of simulated rock section was proposed. Firstly, spatial autocorrelation coefficient was obtained from a group of known distribution grain size of rock section. Secondly, an unknown distribution grain size of rock section was used to calculate spatial autocorrelation coefficient. Finally, linear least squares method about spatial autocorrelation coefficient was solved with constrains. In order to show the feasibility and availability of this method, a serial theoretical rock sections were simulated by random ellipse process. Relationship between spatial autocorrelation coefficient and grain size of simulated rock section was analyzed. With the decrease of offset or increase of grain size of rock section, spatial autocorrelation coefficient is increasing. Grain size distribution of simulated rock section was determined accurately. For example, gritstone was separated into 0.5~1.0 mm, 1.0~1.5 mm and 1.5~2.0 mm, whose computed percentages are respectively 55.8%, 24.6%and 20.2% by the proposed method, closing to the actual values, and the variation trend is same as the actual.

关 键 词：空间自相关系数 粒度分析 最小二乘法 模拟岩石薄片 

分 类 号：P583[天文地球—岩石学]