基于多模态感知的智慧城市环境健康与安全风险预测系统及方法
The smart city environmental health and safety risk prediction system, which utilizes a dual adversarial domain alignment neural network and a three-stream feature fusion network, solves the problem of cross-domain alignment and fusion of multi-source heterogeneous data, achieving efficient and accurate risk prediction and adapting to different scenario requirements.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- BEIJING JINSHANGQI TECH CO LTD
- Filing Date
- 2026-02-11
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies face challenges in processing multi-source heterogeneous data, including significant differences in cross-domain distribution, fragmented data structures, and inconsistent modal characteristics. This leads to difficulties in cross-domain alignment, poor fusion results, weak generalization ability of prediction models, and a lack of flexibility in adapting to different scenarios.
A smart city environmental health and safety risk prediction system based on multimodal perception is adopted. Through heterogeneous data acquisition, domain alignment, joint representation learning and AI large model risk prediction module, cross-domain distribution alignment and modal-level feature calibration are achieved by using a bi-adversarial domain alignment neural network and a weak parameter sharing mechanism. Heterogeneous feature weighted fusion is carried out by combining a time-series-space-graph structure three-stream feature fusion network.
It significantly improves cross-domain alignment, enhances the model's generalization ability and prediction accuracy, has strong adaptability, supports flexible switching in different scenarios, reduces deployment costs, and enhances the model's robustness and anti-interference ability.
Smart Images
Figure CN121860426B_ABST