成像测井图像纹理特征提取及火山岩岩性识别

Extraction of the Imaging Logging Graphic Texture Feature and Volcanic Lithologic Identification

火山岩岩性的准确识别有助于建立区域的地质模型,划分火山的旋回,以及进行详细的地层对比。本文对成像测井所获图像资料进行傅立叶变换,构造频谱图,建立灰度共生矩阵并进行图像纹理特征参数的提取。通过选择提取4种火山岩成像测井图像的4种纹理参数达到对图像纹理的均匀性,粗细度,线性关系,平滑性等的描述,并进而分析纹理参数区分岩性的效果,最终能够提取识别能力最强的纹理描述符,并将获得的已知岩性纹理特征参数数据作为对比资料进行未知火山岩岩性的识别。实际的处理结果显示,成像测井图像纹理提取方法在火山岩识别方面效果较为理想。

Accurate identification of volcanic lithology may help to establish regional geological model,divide volcanic gyration and make detailed stratigraphic correlation.In the paper,Fourier transformation is made to the imaging logging data to establish the spectrogram and greyscale matrix and then extract the texture feature parameters.Through selective extraction of 4 texture parameters of volcanic imaging,the texture uniformity,coarseness degree,linear relationship and smoothness can be described.The most identifiable texture description symbol is extracted to identify volcanic lithology,which will be compared with the unknown volcanic lithology.Actual processed results show that imaging logging image texture extraction is effective in volcanic identification.