An image recognition-based intelligent detection system for defects in automobile valve body machining
The intelligent detection system for automotive valve body machining defects based on image recognition utilizes curvature changes and texture features to filter detection areas, distinguishing between attached burrs and free foreign matter. This solves the problem of poor detection accuracy of valve body bore inner walls, improving detection efficiency and safety.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- TIANJIN JUNLEI PRECISION MACHINERY TECHNOLOGY CO LTD
- Filing Date
- 2026-04-30
- Publication Date
- 2026-07-21
AI Technical Summary
Existing technologies fail to effectively consider the impact of sudden changes in local curvature of the inner wall of automotive valve body bores on defect detection, resulting in poor detection accuracy and potentially causing safety accidents.
An intelligent detection system for automotive valve body machining defects based on image recognition is adopted. The detection area is screened by curvature change characterization value. Combining gray-scale statistical features and contour geometric deviation features, a dual classification decision mechanism is constructed using texture fracture index and shadow offset characterization value to distinguish between attached burrs and free foreign objects.
It improves the accuracy and efficiency of valve body machining defect detection, effectively identifies defects on complex curved surfaces, reduces light and shadow interference, and ensures the reliability and safety of judgment.
Smart Images

Figure CN122435342A_ABST