基于虚拟场景视觉导航的地下管线生成方法、装置和设备
By using a virtual scene visual navigation method, a smooth and continuous 3D pipeline path is generated using policy networks and value networks. This solves the problems of large computational load and non-smooth path in complex environments of traditional algorithms, achieves multi-objective optimization and strong generalization ability, and reduces construction difficulty and cost.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SHANGHAI YINGYI URBAN PLANNINGDESIGN CO LTD
- Filing Date
- 2026-05-18
- Publication Date
- 2026-07-17
AI Technical Summary
Traditional underground pipeline route planning algorithms are computationally intensive in complex 3D spaces, generate uneven paths, lack generalization ability, and struggle to meet multiple design objectives simultaneously, as well as adapt to environmental changes.
A virtual scene-based visual navigation method is adopted. Offline supervised learning is performed through policy network and value network to generate an initial policy model with obstacle avoidance and target approach capabilities. The model is then iteratively optimized under the DDPG learning system to generate a smooth and continuous 3D pipeline path.
The generated pipeline paths are smoother and more natural in three-dimensional space, have strong generalization ability, can adapt to local terrain changes, meet actual engineering needs, reduce construction difficulty and pipe breakage rate, and achieve multi-objective optimization.
Smart Images
Figure CN122199880B_ABST