A grain size auto-classification of Baikouquan Formation, Mahu Depression, Junggar Basin, China

Abstract

The borehole electrical imaging log offers general visual borehole wall, leaving no doubt that the conductivity contrast is sufficient to obtain a qualitative gain-size distribution of rocks. In this study, an automatic grain-size classification method is proposed using gray values of borehole electrical images from Baikouquan Formation in Mahu Depression. The first stage is comparing electrical images with cores. Gravels, sands, silts and clays are all discovered in the cores. The gravels are “mottled” in electrical images, and the bigger the spots, the coarser the gravels. The images of sands are homogeneous bright colored, and the coarser the sandy grains, the brighter the images. The electrical images of silts and clays are homogeneous brown and dark-brown colored. The second stage is auto-discriminating four categories of grain sizes roughly using averages and variances of gray values. The variances of gray values of gravels are high, whereas those of sands are medium. The gray averages of silts are between 160 and 220, whereas those of clays are larger than 200. The third stage is auto-classifying three kinds of gravels or sands finely using frequency distribution of gray values. The gray values of frequency peaks of cobbles are less than 50 and frequencies are larger than 15%, whereas those of pebbles are less than 50 or larger than 200 and frequencies are between 10% and 20%; almost gray frequencies of granules are less than 10%. The dominated gray values of coarse sandstones, medium sandstones and fine sandstones are less than 50, between 50 and 160 and ranged from 160 to 240, respectively. The proposed method is demonstrated to be useful and fast to auto-classify grain size of various rocks in conglomeratic environments.

Keywords

• borehole electrical image
• auto-classify
• grain size
• conglomerate
• baikouquan formation
• mahu depression