中文摘要：

多孔介质微、细观结构的非透明一直是阻碍深入认识深部条件下岩体损伤及煤与瓦斯突出机制的难题。为此，基于小波多分辨分析，提出一种复杂孔隙介质微、细观结构的可视化及多尺度、各向异性的精细描述方法。为验证方法的可行性，对不同地区的6种煤岩试样进行扫描电镜观测；应用小波多尺度变换、图像分割、以及图像重建技术对煤样的SEM图像进行“亚像素”尺度分析；利用小波细节系数重建煤基质的水平、垂直、对角占优的微观孔隙图像；计算微孔的孔喉、孔穴随特征尺度的分布密度，以及微孔与节理的孔隙度与分形维数。用小波多分辨分析方法分割的宏观孔隙具有微米量级，微孔具有纳米尺度。各煤样的宏观裂隙的孔隙度与分维数均大于微观孔隙；孔喉特征尺度分布均具有相似的单峰形式，峰值点决定流体运移的阻力；孔穴的特征尺度分布呈现出单峰、或多峰形态，代表流体的储存能力。研究成果为深入认识复杂孔隙结构对岩体非线性力学行为的影响，提供几何边界、结构参数基于显微图像分析的可行方法。

英文摘要：

Non-transparency of microstructures for the porous media has long been the obstacle for a deep understanding of the mechanisms of rock damage and coal and rock outbursts at depth. As an attempt, a methodology based on wavelet multi-resolution analysis（MRA） for visualization and multiscale-anisotropic-detailed characterization of pore space for complex porous media at meso- and microscopic scales is proposed. In order to validate the proposed approach, observations are made on the six groups of coal rock specimens by using scanning electron microscope（SEM）, sampled from six different underground mining districts; the integrated image processing schemes are conducted for the so called ＂sub-pixel＂ scale analysis of the resulting SEM images, including multi-scale wavelet transform, image segmentation and reconstruction. The detail sub-bands coefficients of the wavelet transform are used to reconstruct images of the horizontal, vertical and diagonal dominant micropores of the microporous matrix. In addition, the probability density distribution of the characteristic length for the pore throat and cavity, the porosity and fractal dimensions for the pores at multiple scales are calculated. The segmented macro-fractures by the MRA are at micron scale while the micro-pores are at nano-scale. The macro-fractures of all the specimens have much larger fractal dimension value than that of the micro-pores. The profiles of the pore throat distribution for all the specimens take on the shape of single peak, and the peak value is the measure of the drag forces for the fluid transport, whereas the curves of the pore cavity take on the shapes of single or multiple peaks, and the maximal peak value represents the storage capacity for the fluids. The obtained results validate the applicability of the microscopic image based approach for characterization of the geometrical boundary and structural features of fractured rocks, which will certainly contribute to a deep understanding of the influences of the complex pore struc