A vegetation phenology dynamic monitoring method based on multi-source data
By processing multi-source data through a multi-head self-attention mechanism and cosine similarity fusion technology, vegetation phenology parameter maps are generated, which solves the problems of accuracy and stability in existing vegetation phenology monitoring and achieves more accurate monitoring of vegetation growth dynamics.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- MUDANJIANG NORMAL UNIV
- Filing Date
- 2026-05-25
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies cannot efficiently and accurately generate vegetation phenological parameter maps, resulting in poor accuracy and stability of vegetation phenological monitoring, especially in terms of collaborative modeling of multi-source heterogeneous remote sensing data and embedding of vegetation growth physical mechanisms.
A multi-head self-attention mechanism is used to perform weighted summation on multi-source data. Combined with vegetation type weight and cosine similarity fusion technology, vegetation phenological parameter maps are generated, including parameters such as growing season length, yellowing period, and greening period.
It improves the accuracy and robustness of vegetation phenology monitoring, enables a better understanding of the intrinsic causal logic of plant growth, generates more accurate phenological parameter maps, and supports ecological impact assessment and management decisions.
Smart Images
Figure CN122412985A_ABST