中文摘要：

针对细编穿刺碳/碳复合材料显微CT细观图像低对比度的特点,提出采用两步法筛选出纤维束和基体之间分界的灰度阈值,将原图像分别转化为纤维束和孔隙的二值图。再结合中值滤波、图像膨胀和腐蚀算法,消除噪声影响、光滑图像边界线。对处理后的细观图像进行统计,发现基体孔隙率满足正态分布,X向和Y向纤维束局部弯曲度分别满足正态分布和Laplace分布。在此基础上建立了包含上述典型细观特征参数的力学模型,预测的拉伸模量和剪切模量满足Laplace分布,泊松比满足正态分布。实验验证了模型预测的拉伸模量分布规律。

英文摘要：

To overcome the difficulty in microstructure distinction of fine weave pierced C/C compos- ites for their low contrast in micro-CT, a two-step filtration was suggested to obtain the eligible grey value threshold indicating fiber bundles/matrix boundary. Binary images of fiber bundles and matrix void were converted from original images, and abatement of noise effect and boundary smoothened by using techniques of mid-filter and inflation/corrosion. Based upon the refined images, matrix porosity and the local curvatures of X fiber bundles are found satisfying normal distribution, while Y fiber bun- dle curvature satisfying Laplace distribution, statistically. A mechanical model which represents these local features is established and the predicted tensile and shear modulus agree with Laplace distribu- tion while the predicted Poisson＇s ratio agrees with normal distribution. The distribution regularity of predicted tensile modulus is validated by experiment.