基于时空卷积网络的草原干旱监测方法
By using a spatiotemporal convolutional network-based approach, multi-source data are fused and meteorological and anthropogenic features are decoupled to generate a true drought monitoring map for grasslands. This solves the problems of misjudgment and threshold adaptability in grassland drought monitoring, and achieves high-precision drought monitoring and ecological management support.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- INNER MONGOLIA NORMAL UNIVERSITY
- Filing Date
- 2025-11-19
- Publication Date
- 2026-07-17
AI Technical Summary
Existing grassland drought monitoring methods cannot effectively distinguish between meteorologically driven drought signals and vegetation degradation caused by human activities, leading to misjudgments and decision-making biases. Furthermore, fixed thresholds cannot adapt to the gradient changes in the intensity of human activities in different grassland areas, and they ignore the differences in the sensitivity of different grassland types to drought.
A spatiotemporal convolutional network-based approach is adopted to dynamically fuse and structure data from remote sensing, meteorology, and human activities, extract meteorological and human vegetation flow features, decouple features using an attention mechanism, perform causal feature fusion and dynamic threshold segmentation, and generate a true grassland drought monitoring map.
It enables accurate extraction and interference-free attribution of drought signals, improves the spatial resolution, temporal accuracy and ecological interpretability of drought monitoring, adapts to the dynamic needs of different grassland ecological zones, and supports ecological management decisions.
Smart Images
Figure CN121302014B_ABST