Method and system for detecting defects of auto-encoder cell panel equipment based on improved yov8

The invention discloses an auto-encoder cell panel equipment defect detection method based on improved yov8, and belongs to the field of electric power detection. The method comprises the following steps: acquiring a plurality of cell panel images; pre-processing the cell panel image to obtain a pre-processed cell panel image; inputting the pre-processed cell panel image into an improved yolov8 model for training to obtain a cell panel surface defect positioning image containing a target frame; performing non-maximum suppression on a target frame in the cell panel surface defect positioning image through an improved NMS algorithm to obtain an accurate target detection frame; and restoring the accurate target detection frame to the corresponding cell panel image to obtain a cell panel surface defect detection result. The invention also discloses an auto-encoder cell panel equipment defect detection system based on the improved yov8. The method can optimize the extracted features for cell panel defect detection