Determining the size of the overlapping area for image stitching in dark-field detection

Dark-field microscopic imaging system has high spatial resolution, but small image field. When an optical surface with a large-diameter is tested, the image stitching is necessary to obtain the full-aperture detection results. In this paper, Harris corner detection algorithm is used to achieve the full aperture testing result by stitching multiple detection images. According to the Shannon sampling theorem, how the size of the overlapping area influences the stitching results is analysed in detail. Combined with the spatial scale of surface defects, a method for determining the size of the overlapping area is given. The standard scratch patterns are used to simulate the stitching process. On this basis, an actual stitching processing is carried out on the detection images of surface scratches of four different spatial scales. Both the simulation and experimental results show that it is reasonable to use the scale of the smallest defect to determine the number of sampling points in the stitching area.

Automated summary

In this paper, the Harris corner detection algorithm is used to achieve the full aperture testing result by stitching multiple detection images. The influence of overlapping area size on the stitching results is analyzed, and a method for determining the size of the overlapping area is given. Both simulation and experimental results confirm that using the scale of the smallest defect to determine the number of sampling points in the stitching area is reasonable.