Classification Model Analysis of Logging Image Based on Metafeatures

In this paper, we proposed a classified strategy on borehole images of logging interpretation. According to the visualized characteristics of borehole images, the analysis of texture is the main objective of this paper. Local Binary pattern (LBP) histogram is adopted to capture the texture information that will be defined as low level textures. Then dominant features of a certain kind will be figured out among low level textures. Finally, after arrangement combination of these dominant features, appropriate meta feature will be presented. The meta feature is a certain kind of feature structure that used as input vectors for classifier.