Core technology
Core technology
Industrial visual recognition system based on deep learning
Construct a textile defect feature enhancement data set and develop a lightweight improvement model.Through migration learning technology, it adapts to the detection needs of different materials, and realizes the real-time identification of 0.1mm yarn defects and oil stains, with a false detection rate of less than 0.5%.
Dynamic tracking and feedback mechanism of whole process data
Deploying a positioning system to establish a database of spatial and temporal trajectories of materials. Develop anomaly propagation simulation model to realize the reverse traceability of quality problems, with a response time of no more than 10 minutes for batch tracing.
Multi-brand/variety seamless switching decision engine
Construct a multi-objective optimization model of production cost, delivery time, and equipment loss, and develop a switching strategy based on Markov's decision-making process.The process parameters are updated synchronously through the Digital Thread technology, and the switching preparation time is reduced by 40%-60%.
Production full-link informationization and intelligent decision-making system
By integrating the four core business systems of ERP, PLM, MES and WMS, the system builds a digitalized closed loop covering order management, process design, production execution and warehousing logistics. Based on the intelligent decision engine, it realizes MRP operation and dynamic scheduling optimization, combines the industry knowledge base with real-time production data, and accurately generates material demand plans and equipment scheduling strategies. Through the whole-link data integration technology, we can open up the end-to-end information flow of supplier access, process parameter monitoring and after-sales quality feedback, forming a closed-loop quality management mechanism of prevention-monitoring-improvement.