Research on Automated Splicing of Optical Surface Defect Images Based on Ar-tificial Intelligence
The quality of automated image stitching using conventional methods is poor.Therefore,an artificial intelli-gence based automated stitching method for optical surface defect images is proposed.The optical component surface defect detection system scales and upsamples the collected images,calculates the image sequence matrix,extracts im-age reference feature points based on the intersection weight of overlapping areas,and uses artificial intelligence algo-rithms to register the images.The reference feature points are matched based on this.By fusing multi frequency band images and analyzing the sensitivity change law of image gradients,the automated stitching area of the image is marked to achieve automatic image stitching.Taking optical surface defect images as the experimental object,the ex-perimental results show that the proposed method obtains automated stitching images with naturalness above 1.0 in different test images.The naturalness of automated stitching images is high,and the quality of automated stitching images is good.