A method for detecting offset of laundry bead sealing thread and real-time correction of hot-pressing position
By constructing a process state fingerprint vector and a localized strategy memory, combined with hash neighborhood retrieval and closed-loop feedback calibration, the problem of response delay and high complexity of existing hot-press sealing correction strategies in multi-variable coupled environments is solved, achieving lightweight, real-time intelligent correction capabilities.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- GUANGZHOU DIFAN DAILY NECESSITIES CO LTD
- Filing Date
- 2026-04-28
- Publication Date
- 2026-06-05
AI Technical Summary
Existing hot-press sealing correction strategies are difficult to achieve real-time adaptation under the influence of multivariate coupling, resulting in inconsistent sealing accuracy. Furthermore, their reliance on complex deep models leads to high deployment complexity and slow response speed, making it difficult to quickly adapt to material conditions and environmental disturbances in industrial settings.
By acquiring data on the temperature gradient of the hot press head, the transient temperature rise rate, and the humidity change of the packaging cavity, and combining this with the output of the image recognition module, a process state input sequence is constructed. An embedded lightweight encoder is used to generate a process state fingerprint vector. A real-time correction strategy is generated by combining a localized strategy memory and a hash neighborhood retrieval mechanism, and the strategy is updated by combining a closed-loop feedback calibration mechanism.
It achieves high-precision error correction within millisecond-level response time, reduces computing power requirements, enhances system robustness and adaptability, supports stable operation of edge devices, and has good interpretability and fast response capabilities.
Smart Images

Figure CN122144269A_ABST