Research On The Key Issues Of Automatic Defects Measurement On Large Fine Optics

Based on the analysis of the existing prototype system for automatic evaluation of defects on the surface of large fine optics, this paper proposed an improved defect evaluation solution for enlarging the aperture range of detection. By improving the three important aspects of the prototype system including illumination system, image mosaic algorithm, image data storage, the integral performance of the system was enhanced, and the applicability of evaluating a great amount of large fine optics in the practical situation was promoted.In the aspect of illumination, the deficiencies in illumination uniformity and structural compactness of the previous illumination system based on fiber quartz halogen lamp and critical illumination was analyzed, then an improved illumination structure based on LED and Kohler illumination was proposed. Through the design and simulation with optical software and instrumentation of the new structure, the improvement on illumination area, uniformity and compactness of the new system design was proved. Then a further improved structure based on LED and fly-eye lens was presented. The design and simulation proved that illumination area and uniformity was further improved without discarding the compactness.In the aspect of sub-aperture image mosaic, a multi-cycle image mosaic algorithm based on feature classification was proposed after analysis on the advantages and disadvantages of the previous edge-expansion method. The algorithm was constructed by two parts: feature classification and multi-cycle image mosaic. The former part divided overlapped areas of the sub-aperture images into four types based on extraction and classification of features in overlapped areas. Then the latter part orderly conducted four cycles of image mosaic to get the final panoramic image according to different situations of the four types of overlapped areas. Experiment result showed that the algorithm could effectively avoid mosaic mismatch caused by locating error accumulation of guide rail and run-through line features and the efficiency was promoted.In the aspect of image data storage, existing methods and standards of the image compression was analyzed, then a solution for gray image storage base on JPEG2000 was proposed. The experiment showed that the solution not only made efficient use of storage space, but also had quite a large flexibility. Then a structured storage solution using Segment encoding for binary image produced in defect evaluation was proposed after analyzing demand of the defect evaluation system on binary image storage. The experiment showed that this solution could not only save the storage space, but also implement random read and reconstruction in defect feature level.Finally, an integral experiment in system level was launched. The results further demonstrated the integral system promotion due to the improvement on the above three aspects and applicability in practical detecting situation. In addition, the lower bounds of the quality factor of image compression in two evaluation accuracy levels were found after the analysis on the influence of lossy image compression on defect evaluation of the system.