An AI vision-based feed production process real-time monitoring and tracking system
By using AI vision technology to monitor and track feed production in real time, combined with an improved GLNet network and dense optical flow algorithm, the shortcomings of existing technologies in status recognition and batch traceability have been solved, achieving stable recognition and accurate traceability, and improving the ability to make timely decisions and control risks in production management.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- BOXING COUNTY GENQING AGRICULTURE & ANIMAL HUSBANDRY TECHNOLOGY CO LTD
- Filing Date
- 2026-03-09
- Publication Date
- 2026-06-16
AI Technical Summary
The current feed production process lacks real-time status monitoring and batch traceability capabilities, resulting in delayed anomaly detection, difficulty in locating anomalies, and problems such as batch mixing and mis-binding. It is difficult to achieve continuous and stable status identification and data alignment, which affects product quality stability and production safety.
A real-time monitoring system based on AI vision is adopted, which combines an improved GLNet network and dense optical flow algorithm. Through video acquisition, preprocessing, abnormal area localization and motion causal reasoning, a set of abnormal process events is generated and bound to a batch event chain to achieve stable identification and accurate traceability.
It achieves stability in status recognition and timeliness in anomaly detection under complex operating conditions, reduces the risk of batch mixing and misbinding, and improves the ability to trace quality and collect audit evidence throughout the entire process.
Smart Images
Figure CN122222933A_ABST