According to ISO10110-7 surface defects standard, a super-smooth surface defects measurement and evaluation system is established. High quality defects image is acquired by a specific microscopic dark-field scattering imaging device. Combined with a precise XY-scanning system for sub-scanning the optics surface, panoramic surface defects image could be obtained. Surface defects imaging model is discussed in view of scattering. In the new system, LED is used as the emitting source to realize system illumination. Compared with the halogen tungsten lamps, the new light source has less weight and smaller size. According to the practical situations of defects images, probable mismatches of template matching method of original system are analyzed, a feature based multi-cycle image mosaic algorithm is developed to mosaic sub-aperture defects images. In the new algorithm, images are mosaicked in an order which is decided by the results of feature extraction in their overlapping areas. Feature extraction is firstly carried out to recognize defects features in the overlapping areas of images. Images with defects in their overlapping areas are mosaicked in the way of region growing to form many image blocks. Then images and image blocks with no defects in overlapping areas will be combined into panoramic image according to the nominal displacement distance of the stage. A complete set of image preprocessing, recognizing and evaluating is established to extract defects features and calibrate. With this system, reliable recognition, pinpoint and accurate evaluation of surface defects could be rapidly, automatically realized, with a resolution reaching micron-level.

Date Published: 11 October 2010
PDF: 5 pages
Proc. SPIE 7656, 5th International Symposium on Advanced Optical Manufacturing and Testing Technologies: Optical Test and Measurement Technology and Equipment, 76560A (11 October 2010); doi: 10.1117/12.867072

Published in SPIE Proceedings Vol. 7656:
5th International Symposium on Advanced Optical Manufacturing and Testing Technologies: Optical Test and Measurement Technology and Equipment
Yudong Zhang; Jose M. Sasian; Libin Xiang; Sandy To, Editor(s)