Visual inversion method for surface oxide thickness of an automobile die-casting mold steel forging
By employing multi-channel image processing and self-calibration mapping, the inaccuracy and instability of oxide scale thickness detection in existing technologies have been resolved, enabling efficient and accurate detection of oxide scale thickness on the surface of automotive die-casting mold steel forgings.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- ZHEJIANG LIYUAN ZHONGGONG SCI & TECH CO LTD
- Filing Date
- 2026-06-12
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies for detecting the thickness of oxide scale on the surface of automotive die-casting mold steel forgings suffer from problems such as reliance on manual experience, destructive testing, high equipment costs, high requirements for environmental stability, unstable image features, easy failure of calibration relationships, and failure to fully utilize natural cross-sectional references, resulting in inaccurate and unstable test results.
By acquiring multi-channel images, marking surface boundaries to generate effective pixel areas, identifying the lateral layer of oxide scale using natural cross-sectional edges, establishing a self-calibrated mapping relationship, and combining gradient calculation and edge judgment to generate oxide scale thickness inversion values, selecting thickness samples and constructing visual feature vectors for thickness inversion.
It enables accurate and stable detection of oxide scale thickness without relying on external calibration or human experience, reducing equipment costs and environmental dependence, improving the standardization and reliability of detection, and adapting to different forgings and environmental changes.
Smart Images
Figure CN122415598A_ABST