Defect detection of mobile phone screen based on improved difference image method

The defect inspection with machine vision plays an important role in the quality control of mobile phone screen manufacturing. An improved difference image method is proposed to detect the defects of mobile phone screen. Considering the problem of misalignment of mobile phone screen image, PatMax algorithm and image correction technology based on affine transform are adopted to realize pixel alignment between template image and testing image. With the intention of eliminating noise effects and protecting the edge information of the test image, a image filtering is performed by 3×3 median filtering. Based on the above image preprocessing results, the defect-free image is used as the template image and the fuzzy processing is carried out, the differential operational between the testing image and the template image is performed to obtain the residual image, and the improved Otsu dual threshold method is used to achieve defect judgment on residual images. The experimental results show that the approach proposed in this study has achieved high efficiency and accuracy.