基于X射线CT的混凝土内部裂纹识别研究

Identification of crack in concrete interior based on X-ray CT

利用医用CT设备采集了完整单轴压缩作用下的混凝土断面图像,分别采用阈值分割、聚类和图像差值三种算法提取了混凝土内部裂纹,讨论了三种算法提取混凝土内部裂纹时各自的优缺点,并对裂纹提取结果进行分析和评价。结果表明:阈值分割和聚类算法均难以区分裂纹和孔隙,图像差值算法不仅可以较好地提取裂纹,还能区分裂纹与孔隙;阈值分割和图像差值算法均需要通过大量试验确定合理阈值,聚类算法不需要设定阈值,可以避免选择阈值导致的人为误差;基于差值算法得到的差值图像可以定量描述裂纹的演化过程;随着CT设备的不断发展,图像差值算法将在混凝土细观结构研究中应用越来越广泛。

Collect a series of section images of concrete under uniaxial compression by using medical CT imaging devices,and then extract internal cracks in concrete by threshold segmentation method,clustering algorithms method and difference algorithm method.And discuss the advantages and disadvantages of these three algorithms when applicated to extract concrete internal crack.At last,analyze and evaluat the results of crack of extraction.The results shows that it is hardly to distinguish cracks from gaps by threshold segmentation method or clustering algorithms method,but it is easy to do so by difference algorithm method.And it is easy to identify cracks by difference algorithm method.Both using threshold segmentation method and difference algorithm method need a large number of test in order to set a reasonable threshold,but using clustering algorithms method needn＇t.So it can avoid error of threshold value.The Images based on difference algorithm method can describe the evolution of crack quantitatively.With the development of CT equipment,difference algorithm method will be applied to the microstructural study more widely.