State-of-the-Art on Texture-Based Well Logging Image Classification

Imaging well logging can easily and effectively determine the existing locations of the reservoirs, because of its ability of showing the stratal lithologies and geometric changes in the form of image. Recently, imaging log is used widely for its higher discernibility and has become a research hotspot in the field of well logging technology. How to fully utilize image processing, pattern recognition and other related theoretical methods for both more precise quantitative evaluation of imaging logging and the interpretation of images is the focus. Texture analysis plays an important role in the field of computer vision and pattern recognition. This paper surveys the research background and current situation of the imaging well logging, and then reviews most existing typical algorithms for texture analysis. It focuses on the grey level co-occurrence matrices (GLCM) algorithm, local binary patterns algorithm (LBP), Gabor transform, wavelet transform, as well as Contourlet transform, and analyzes their respective pros and cons. Based on this, and considering the features of the logging images, this paper gives a method for the classification of the well logging images, and proposes a system model of logging image recognition and classification. The problems, prospects for future development and suggestions for further research works are put forward at the end of the paper.