Multiview AOI System for Automotive Display Panels Uses Improved YOLOv8n Detector
A study reports a multiview automated optical inspection system for automotive display panels, built around YOLO-OG, a detector derived from YOLOv8n. The authors say it combines multiview imaging, dust suppression and region-of-interest preprocessing, and report gains over the baseline and in production-line validation.
Researchers have developed a multiview automated optical inspection system aimed at surface defects on automotive display panels. The work proposes YOLO-OG, a detector built on YOLOv8n for use in production-line conditions.
The workflow combines image capture from multiple views, dark-field dust suppression, region-of-interest preprocessing and defect recognition. The authors report that YOLO-OG improved precision and mean average precision over the YOLOv8n baseline while keeping detection efficiency comparable.
In production-line validation, the authors report higher panel-level recall than manual inspection and fewer false positives than conventional AOI under the same imaging workflow. They also report transfer experiments in which the gains held after fine-tuning on the PKU MSD dataset.
For equipment buyers, ask whether the multiview setup fits your panel sizes and line speed, and how dust and reflection control would be handled on your line. Request validation data from panels like yours, and clarify how the detector would be tuned and maintained after installation.