Design of automatic detection algorithm for dislocation contrasts in birefringence images of SiC wafers

Birefringence imaging is one of the powerful methods for non-destructive characterization of defects in the semiconductor crystals. However, due to the complicated and unclear contrasts of dislocations in the birefringence image, it was considered to be difficult to automatically detect the position of the dislocation contrasts by the conventional image processing. In the present study, we designed the automatic detection algorithm for the dislocation contrasts taking into account the characteristic feature of the dislocation contrasts, which were always pair of black and white contrasts. To detect the large change in the contrast level near the dislocation contrast, the automatic detection algorithm was constructed by using a variance filter. Finally, we succeeded in detecting the position of the dislocation contrasts with relatively high precision and sensitivity.

Automated summary

An automatic detection algorithm was designed to detect the positions of dislocation contrasts in semiconductor crystals, utilizing the characteristic feature of black and white contrasts of the dislocations and a variance filter to identify large changes in contrast levels. The algorithm successfully achieves a relatively high precision and sensitivity in detecting the dislocation contrasts.