基于Tamura纹理和广义判别分析的岩屑图像识别算法

自动、快速、客观地对岩屑进行分类是录井过程中一个非常重要的问题.以往在现场主要是靠人工经验的方式对各类岩屑的百分含量进行估测.但该方法受人为因素的影响较大.本文针对这个问题,提出基于Tamura纹理和广义判别分析的岩屑图像识别算法.算法采用Tamura纹理特征对不同岩性的岩屑图像进行纹理提取,然后将提取的Tamura纹理特征带入GDA(Generalized Discriminant Aanalysis)分类器进行岩屑的识别.通过对四川某地气田4口井的现场随钻的实验结果表明本文算法识别算法具有较高的识别率,为快速自动岩屑分类提供了一个有效的方法.

An algorithm based on tamura texture and GDA for rock cutting recognition

In the cutting logging process, automatic, rapid and objective classification of cuttings is a very important. The past, artificial experience is the main way to to estimate the percentage of different kind of cuttings. However, the method is not very objective. According the problem, in this paper an algorithm based on Tamura texture and GDA is proposed for cutting recognition. First, the Tamura texture features of cuttings are obtained. Then the features are used to train the GDA classifier. Finally, the cuttings are classified by using the trained GDA. Using the algorithm, after logged 4 wells of Sichuan gas reservoir, a high recognition rate can be reached. The experimental results demonstrate the algorithm is feasible, robust and applicable.