Metal bottle cap surface electroplating defect traceability analysis method for micro defects
By constructing a process anchoring module and a temporal causal mask generator, and combining multi-source heterogeneous data and a historical case library, the problems of dynamic tracing difficulties and high misjudgment rates in the source analysis of electroplating defects in metal bottle caps are solved. This enables efficient and interpretable defect cause analysis, which is suitable for real-time optimization in industrial settings.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- DONGGUAN YIHAN HARDWARE PRODUCTS CO LTD
- Filing Date
- 2026-03-20
- Publication Date
- 2026-06-16
AI Technical Summary
Existing technologies for tracing and analyzing the sources of electroplating defects on metal bottle caps suffer from problems such as difficulty in dynamic tracing, lack of interpretability, poor model flexibility, high misjudgment rate, and high deployment cost. In particular, in scenarios with multiple processes and long time spans, it is difficult to achieve accurate defect cause analysis.
By acquiring the timestamps, equipment identifiers, and key process parameters of metal bottle caps in the pre-processes of stamping, cleaning, and passivation, a process anchoring data sequence carrying multi-dimensional working condition characteristics is generated. Combined with high-resolution defect images for weak supervision alignment, a spatiotemporally aligned multi-source heterogeneous fusion dataset is generated. Causal inference is performed using a temporal causal mask matrix, and confidence is verified by combining a historical case library. Finally, a structured causal tracing report is generated.
It achieves high efficiency, interpretability, and reliability in cross-process defect tracing, reduces computational overhead, is suitable for the real-time and safety requirements of industrial sites, and provides efficient quality optimization support.
Smart Images

Figure 1 
Figure 2