Sub-aperture stitching strategy for dark-field imaging of large fine optical components

Scratches, digs, and other defects play an important role in the quality control of fine large-aperture optical components. Dark-field microscopic imaging has become one of the most common methods for surface defect detection. Nevertheless, compared to the significant increase in the aperture of the test component, the imaging field of view is still very limited. Therefore, the sub-aperture stitching strategy can expand the detection range dynamically without reducing the resolution in the detection of large optical components. Sub-aperture images usually are matched at adjacent positions by feature matching. Nevertheless, there may exist contradictions in the feature matching of overlapping areas. Also, some sub-aperture images only have defects in nonoverlapping areas, which cannot be solved by feature matching, resulting in inaccurate defect localization. In this paper, a linear constraint sub-aperture (LCSA) stitching strategy is proposed. The results of feature matching are converted into the linear constraints of all step errors on the scanning path, and the optimal solution of the step errors is obtained through least-square optimization. As a result, high-precision global stitching can be realized by correcting the step errors. In addition, the mean square error (MSE) based on the feature matching results is proposed to evaluate the stitching results. Experimental results demonstrate that this strategy can reduce the MSE to 3.4%-13.6% of the direct stitching and has strong robustness under different experimental conditions. Herein, the quantitative matching results as feature-level information are employed for global optimization, which makes up for the lack of local defect localization accuracy of the feature matching algorithm. It also helps mitigate the limitations of a few matching features and can improve the overall reliability for defect detection of large fine optical components.