Dot sorting inspection method, device, equipment and storage medium
By integrating multi-source data and intelligent analysis, abnormal behaviors are identified and collaborative work orders are generated, which solves the problems of single detection and inefficient processes in existing technologies, and achieves efficient risk warning and closed-loop management.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SHANGHAI YUNDA HIGH TECH CO LTD
- Filing Date
- 2026-02-11
- Publication Date
- 2026-06-09
AI Technical Summary
Existing sorting and inspection methods at network outlets are limited in their detection capabilities, lack accurate identification, predictive ability, and efficiency, making it difficult to achieve standardized closed-loop management.
By acquiring and standardizing delivery and transportation data, basic information of service points, and monitoring visual data from multiple data sources, feature extraction and fusion are performed using convolutional neural networks and Transformer encoders. Anomaly risk prediction is then performed using the XGBoost model to generate collaborative work orders and track their status.
It improved the accuracy of anomaly identification, enabled risk warning and closed-loop management, and enhanced the management efficiency and standardization of the sorting operation site.
Smart Images
Figure CN122175353A_ABST