一种基于深度视觉的漆包线线轴表面缺陷在线检测方法
By utilizing depth vision technology and polarization features and frequency domain transformation processing, the problem of low feature discrimination in the detection of surface defects of enameled wire spools has been solved, enabling efficient identification and accurate positioning of multiple types of defects, and improving the accuracy and sensitivity of detection.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHANGZHOU WELLYUN ELECTRICAL
- Filing Date
- 2026-06-18
- Publication Date
- 2026-07-17
AI Technical Summary
In the detection of surface defects of enameled wire spools, existing technologies cannot decouple physical morphological changes from optical appearance changes of the enamel film using ordinary intensity imaging. This results in low distinguishability of defect features such as scratches and oxidation, and color analysis is not sensitive to early color changes, making it difficult to accurately identify multiple types of defects.
A depth vision-based approach is adopted to acquire pseudo-straight line circumferential images by configuring a linear array camera with orthogonal double polarizers. By combining polarization feature calculation and frequency domain transformation, polarization degree channel images and total light intensity channel images are generated. Axial curvature maps and chromaticity deviation distribution fields are extracted. Defect features are fused using mutual verification coupling factors to determine the type and location of defects.
It improves the geometric stability and positioning accuracy of defects on the outer cylindrical surface of spools, enhances the detection sensitivity of small morphological defects, expands the coverage of detectable defect types, and improves the discrimination accuracy in scenarios where multiple types of defects coexist.
Smart Images
Figure CN122409696A_ABST