发泡混凝土孔结构快速检测与精准识别

Rapid Detection and Accurate Identification of Foamed Concrete Pore Structure

摘要:

提出了基于零度光源的非接触孔结构检测和识别方法，用零度光源辅助采集孔结构图像信息，并改进了图像降噪的均值滤波算法，研究了孔结构特征识别方法.结果表明：零度光源能大幅增强发泡混凝土孔隙图像的对比衬度，节约样品后处理时间；改进均值滤波法可有效抑制孔隙图像噪声，提升孔隙边缘的清晰度；优选Canny算子增加了薄壁孔结构特征识别的精准性.该方法实现了发泡混凝土孔边缘特征、孔径分布和孔隙率等孔结构信息的快速检测与精准识别，有望用于其他发泡材料的孔结构表征.

Abstract:

A non-contact pore structure detection and identification method based on zero-degree light source was proposed. The pore structure image was acquired with the assistance of the zero-degree light source， and the mean filtering algorithm for image denoising was improved. The pore structure feature identification method was studied. The results show that the zero-degree light source can greatly enhance the contrast of the pore image of foamed concrete， thus the post-processing time is saved. The improved mean filtering method can effectively suppress the noise of pore image and improve the clarity of pore edge. The optimized Canny operator increases the accuracy of structural feature identification of thin-walled pores. The rapid detection and accurate identification of pore structure information are realized by this method， such as pore edge characteristics， pore size distribution and porosity of foamed concrete， which is expected to be used for pore structure characterization of other foamed materials.