Design of a real-time abnormal detection system for rotating machinery based on YOLOv8

View Full Paper

Ask AI

Bookmark

The detection system achieves an average accuracy of 97.8%, showcasing its effectiveness in real-time abnormal detection.
Utilizing YOLOv8 with depthwise separable convolution enhances detection efficiency for small abnormalities in complex conditions.
The method employs a temporal motion compensation module to correct vibration displacement between frames for improved accuracy.
Deployed on the Jetson AGX Xavier platform, the system maintains a processing speed of 29.5 FPS, suitable for industrial applications.

Discussion

No takes yet. Share an insight, caveat, or question.

Proposed system achieves 97.8% accuracy in detecting abnormalities in rotating machinery, indicating strong real-time capabilities.

Chen et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark