针对复杂纹理的手机外壳缺陷检测方法

A DEFECT DETECTION METHOD FOR COMPLEX TEXTURE ON MOBILE PHONE SHELL

针对手机表面存在纹理复杂、缺陷细小和缺陷多样性等特点,提出一种应用于复杂纹理的手机外壳缺陷检测方法。该方法首先对手机外壳表面划分区域进行放大处理,并抑制放大后的纹理。然后进行图像的主结构提取,再采用一种自适应边缘检测Roberts方法来进行缺陷轮廓提取。最后运用一系列的中值滤波、形态学运算等方法来综合处理并提取缺陷。实验表明,对手机外壳的划痕、污渍等缺陷误检率低,具有良好的检测效果。

Considering the features of complex texture, tiny defect and defect diversity of the surface of mobile phone shell, we propose a mobile phone shell defect detecting algorithm based on a complex texture. This method firstly magnified the partition area of the mobile phone shell, and suppressed the main texture of it. Then, the main structure of the image was extracted and self-adaption Roberts edge detection method was used to extract the defect contour. Finally we integrated handled and extracted the defect according to a series of median filter, mathematical morphology operation, etc. The experiments show that this method can provides good results for detecting the scratch, stain and other detects on the surface of mobile shell, with a low error rate.