A dynamic control method and system for shrimp freezing process
By using a knowledge graph generated through a lightweight instance segmentation network and a knowledge distillation algorithm, combined with a neural symbolic reasoning system, the shrimp freezing process is dynamically controlled. This solves the problems of slow response speed and poor adaptability in existing technologies, achieving efficient and interpretable shrimp freezing control, and improving freezing quality and system adaptability.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- PANJIN MEIRI GRP CO LTD
- Filing Date
- 2025-11-19
- Publication Date
- 2026-04-10
AI Technical Summary
Existing technologies for shrimp freezing suffer from slow response speed, lack of interpretability in decision-making, limited computing resources at edge nodes, and poor system adaptability, resulting in low control precision and difficulty in quality protection during the freezing of damaged shrimp.
A lightweight instance segmentation network is used to extract damage feature vectors, which are then combined with a knowledge distillation algorithm to generate a lightweight knowledge graph. This graph is deployed on edge computing nodes and dynamically controlled through a neural symbolic reasoning system to generate differentiated freezing control parameters. Finally, it is optimized online through a differentiable reasoning mechanism.
It achieves millisecond-level decision-making speed, provides interpretable decision-making basis, enables differentiated control for different damage types and severity, and improves freezing quality and system adaptability.
Smart Images

Figure CN121168672B_ABST
Abstract
Citation Information
Patent Citations
Flood control emergency plan generation method, system and equipment and storage medium
CN121093173A