Remote monitoring method and system for super-deep underground space based on multi-source data fusion
By combining interface specifications, conversion modules, clustering algorithms, and prediction models, the problem of integrating multi-source heterogeneous data was solved, enabling real-time and accurate monitoring of ultra-deep underground spaces and improving risk warning capabilities.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SINOHYDRO BUREAU 8 CO LTD
- Filing Date
- 2025-11-06
- Publication Date
- 2026-07-21
AI Technical Summary
Existing monitoring technologies struggle to effectively integrate multi-source heterogeneous data and geospatial information, resulting in low efficiency in data collection and analysis for ultra-deep underground space monitoring, and an inability to accurately reflect the actual state of underground space.
Initial data is obtained through a pre-defined interface specification. A conversion module is used to handle format differences. A clustering algorithm is used to group similar devices, embed location information, and generate an integrated dataset with annotations. Combined with the correlation information of environmental variable changes, the dataset is fused into a dynamic matrix. A classification model is used to identify abnormal patterns, generate a mapping layer, and a prediction model is used to estimate trend information to generate a risk distribution map.
It has achieved efficient collection and transmission of multi-source data, accurately mapped the actual state of underground space, constructed a real-time and accurate risk early warning system, and significantly improved the efficiency of environmental monitoring response.
Smart Images

Figure CN121524916B_ABST