Design and Research on Visual Detection System for Cell Number of Honeycomb Ceramics

Honeycomb ceramics are commonly used as catalyst carriers, whose cell number is a crucial criterion for evaluating catalytic performance. In industrial production, the cell number of honeycomb ceramics is primarily measured manually, resulting in low efficiency and time consumption. As for these problems, a machine vision-based detection system is designed. In this system, the end face images of honeycomb ceramics are preprocessed by a designed hybrid filtering method to obtain through-hole number; an automatic dual-threshold Canny algorithm is designed to detect the sub-pixel outer circle of the end-face contour by combining sub-pixel contour extraction and least-squares fitting to compute the end face area and the cell number. In the experiment, the system's detection results are compared with relevant standards, revealing that the experimental samples had an average diameter deviation of -0.9689 mm and an average cell number deviation of - 2.39879 cells/cm²; when compared to manual measurements, the system showed a diameter difference of 0.8113 mm and a cell number difference of 0.46385 cells/cm², demonstrating lower error; and the system consumes an average time of 0.6494s. These results indicate that the system can achieve accurate, stable, and efficient cell number detection of honeycomb ceramics, offering significant value for industrial production.