Grinding wheel mesh automatic detection method based on improved SwinTransformer
By improving the SwinTransformer network structure and feature fusion module, the problems of small target information loss and poor multi-scale feature fusion in grinding wheel mesh detection are solved, achieving high-precision and robust automatic detection, which is suitable for complex industrial environments.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- HARBIN INST OF TECH ZHENGZHOU RES INST
- Filing Date
- 2026-04-27
- Publication Date
- 2026-07-17
AI Technical Summary
Existing deep learning methods are difficult to effectively capture detailed information and complex texture features of small targets in the detection of abrasive wheel mesh, and lack robustness in variable environments, resulting in low detection accuracy and high false negative rate, which cannot meet the high precision requirements of industrial environments.
An improved SwingTransformer network structure is adopted, which combines image enhancement technology and feature fusion module, including up-merging module and skip connection strategy, to improve feature extraction and multi-scale fusion capabilities, and enhance the adaptability and robustness of the model.
It significantly improves the accuracy and robustness of grinding wheel mesh inspection, enabling efficient and accurate detection of targets of different sizes, angles and lighting conditions in complex industrial environments, and enhancing the model's adaptability to changing environments.
Smart Images
Figure CN122415552A_ABST