陶瓷膜表面缺陷的表征与分类研究

Characterization and Classification of Surface Defects in Ceramic Films

陶瓷膜种类多而杂,对陶瓷膜缺陷的检测则涉及到其表面的缺陷检测(支撑体层),而机器视觉的发展和运用使表面缺陷检测变得更加简单且智能化.文章基于陶瓷膜表面存在的划痕、裂纹、落渣、凹坑4种缺陷,运用MATLAB的图像处理技术和BP神经网络分类对陶瓷膜的表面缺陷进行提取分析和分类,结果表明,BP神经网络分类对陶瓷膜表面缺陷识别的正确率达到78.125%.

The types of ceramic films are numerous and complex.The defect detection of them involves the surface detection on the support layer,which becomes easier and more intelligent based on the development and application of machine vision.In this paper,MATLAB image processing techniques and BP neural network classification method were used to detect,analyze and classify the surface defects of ceramic films according to the scratches,cracks,dregs,and pit.The experiments show that the accuracy of the surface defects of the ceramic film by BP neural network reached 78.125%.