Adaptive-reflective-based tool-augmented imaging optical system automatic design method
By constructing a hybrid knowledge base in the field of optics and an adaptive reflective thinking chain, combined with a reinforcement learning training process, the problems of dependence on initial structure and local minima in the design of imaging optical systems are solved, and efficient and interpretable automated design is achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHANGCHUN INST OF OPTICS FINE MECHANICS & PHYSICS CHINESE ACAD OF SCI
- Filing Date
- 2026-06-11
- Publication Date
- 2026-07-17
AI Technical Summary
Existing imaging optical system design methods suffer from problems such as strong dependence on initial structure, susceptibility to local minima, lack of physical interpretation and external knowledge utilization, resulting in low design efficiency and poor imaging quality.
We employ an adaptive reflection-based tool enhancement approach, which constructs a hybrid knowledge base in the field of optics, defines a reflective thought chain, and combines adaptive hybrid knowledge routing and reinforcement learning training processes to achieve self-reflection and global optimization of the agent.
It improves the global optimization capability of imaging optical system design, shortens search time, enhances the interpretability and efficiency of design, ensures the accuracy and reliability of knowledge, and realizes efficient automated design.
Smart Images
Figure CN122413982A_ABST