基于数字图像处理的混凝土微裂纹定量分析技术

Automatic Quantitative Analysis of Microcracks in Concrete Based on Digital Image Processing Techniques

在应用真空荧光环氧浸渍技术得到含微裂纹的混凝土显微图像的基础上,基于MATLAB搭建开发了由二值化、形状分析、骨架化和自动测量4个模块组成的混凝土微裂纹定量分析软件QUANSIMIC(quantitative analysis system of microcracks in concrete),可实现微裂纹区域的识别、提取以及面积、长度及宽度等二维结构特征参数的测量分析.其中二值化模块中采用Intermodes法确定灰度阈值进行微裂纹区域的识别和提取;形状分析模块中分别采用开操作去除噪声,采用长宽比为2.0的判别准则区分孔隙与裂纹;在自动测量模块中的长度计算时考虑了微裂纹骨架的宏观倾角,相对误差很低.

A MATLAB-based software package called QUANSMIC（quantitative analysis system of microc- racks in concrete） including image-binarizing module, shape-analyzing module, crack-skeletonizing module and automatic measuring module was developed to automatically extract and quantitatively analyze the mi- crocrack characteristics（such as area, length, width, etc） in the microscopic images of concrete slices, in which the microcracks were highlighted with vacuum-fluorescent-epoxy-impregnation technique. The In- termodes method was employed to determine the grayscale threshold in the binarizing module. The noises were removed by opening operation and the aspect ratio（ratio of length to width） was set at 2.0 to distin- guish voids from cracks in the shape-analyzing module. The macro-orientation of the skeletonized microc- racks was considered in length measurement in the automatic-measuring module to reduce relative errors. QUANSMIC could provide a powerful tool of quantitatively damage evaluation and health diagnosis of con- crete based on digital-image-processing techniques.