基于LBP和GLCM的煤岩图像特征提取与识别方法

Coal Rock Image Feature Extraction and Recognition Method Based on LBP and GLCM

针对煤矿开采工作面无人化要求,提出一种基于LBP和GLCM的煤岩图像特征提取与识别方法。采用LBP算法判断煤块与岩石纹理存在差异性,然后通过GLCM实现煤块与岩石图像在水平、直角、45°、135°方向上的灰度共生矩阵,并完成对能量、熵值、对比度、逆差分矩等4个煤岩图像纹理特征参数提取。试验表明:LBP算法在检测煤块与岩石局部纹理特征差异的过程中,具有一定的高效性,但存在不足,后续通过GLCM提取的煤岩图像特征参数,可以找到适用于煤岩分类的特征参数,增加煤岩识别的鲁棒性。

Aiming at the unmanned requirements of coal mining face, a method for extracting and identifying coal rock image features based on LBP and GLCM is proposed. The LBP algorithm is used to judge the difference of the rock texture of the coal block. Then, the GLCM is used to realize the gray level co-occurrence matrix of the coal block rock image in the horizontal, right angle, 45 degrees and 135 degrees directions, and the energy, entropy value, contrast and inverse difference moment are completed. Extraction of texture feature parameters of four coal rock images such as partial moment. Experiments show that the LBP algorithm has certain efficiency in detecting the difference of local texture characteristics between coal and rock, but there are some shortcomings. The characteristic parameters of coal and rock image extracted by GLCM can be found to find the characteristic parameters suitable for coal and rock classification to increase the robustness of coal rock identification.