Fig. 1. Principle of photoluminescence imaging

Fig. 2. Camera layout and image view (10×6 component)

Fig. 3. Stitching matches key points of Camera 2

Fig. 4. Single camera projection positioning

Fig. 5. Neighborhood extreme value difference processing of camera Y-direction projection curve

Fig. 6. Edge index of camera

Fig. 7. Movement direction of each camera

Fig. 8. Assembly result of component image

Fig. 9. Single cell plate front and PL image

Fig. 10. Diagrams of corner defects

Fig. 11. Location of corner points to be detected

Fig. 12. Corner screenshot process

Fig. 13. Convolutional neural network models for training

Fig. 14. Part of training samples

Fig. 15. Model training accuracy and loss curve