微型磁瓦表面线缺陷视觉检测方法研究

针对微型磁瓦图像对比度低、纹理背景复杂、亮 度不均匀、缺陷区域小等特点,本文提出了一种微 型磁瓦表面线缺陷视觉检测方法。首先,构造自适应静态掩膜,屏蔽微型磁瓦轮廓；其次, 采用非线性各 向异性扩散方程,抑制微型磁瓦表面纹理；进而,构造动态掩膜自下而上扫描磁瓦图像,提 取线缺陷区域； 最后,在开发的微型磁瓦视觉检测实验装置上,进行了大量的实验研究。实验结果表明,本 文缺陷提取算 法能够较准确提取出磁瓦表面图像的线缺陷,表面缺陷检测的准确率为94.6%。

Study on visual detection method of surface linear defects on micro-magnetic ti le

The micro-magnetic tile image has the characteristics of low contrast ,complex texture background, uneven brightness and small defect area.This paper proposes a method of linear detects on micro-magnetic tile surface.Firstly,an adaptive static mask is constructed to shield the micro-ma gnetic tile profile.Secondly,the nonlinear anisotropic diffusion equation is used to suppress the surface texture of the micro-magnetic tile. Further,the dynamic mask is constructed to scan the magnetic tile image from th e bottom to the top,and the linear defects are extracted.Finally,an experimental device for visual inspect ion of micro-magnetic tiles was developed and a large number of experimental studies were carried out.The exper imental results show that the defect extraction algorithm can accurately extract the linear defects on micro- magnetic tile,and the accuracy of surface defect detection is 94.6%.