A method for spatiotemporal prediction of satellite cloud track from cloud images
By using the 3D-ATAM-BiGRU spatiotemporal prediction model and physical constraint loss function, the shortcomings of existing cloud map trajectory prediction methods in terms of spatiotemporal feature extraction and physical consistency are addressed, achieving high-precision cloud movement trajectory prediction and improving the stability of meteorological monitoring and solar power generation.
CN122115507APending Publication Date: 2026-05-29NANJING UNIV OF INFORMATION SCI & TECH
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- NANJING UNIV OF INFORMATION SCI & TECH
- Filing Date
- 2026-04-29
- Publication Date
- 2026-05-29
Smart Images
Figure CN122115507A_ABST
Abstract
The application discloses a kind of satellite cloud picture cloud track space-time prediction method, belong to meteorological prediction and satellite remote sensing image processing technical field, including using sliding window strategy from preprocessed satellite cloud picture L2 time series data intercept fixed length input sequence and its corresponding target output sequence, build paired training sample;Establish physical constraint loss function;3D-ATAM-BiGRU space-time prediction model is constructed, the model is trained using physical constraint loss function, preprocessed satellite cloud picture L2 time series data is input into the 3D-ATAM-BiGRU space-time prediction model after training, generates future multi-time step cloud physical parameter prediction result.The application can effectively capture the space-time characteristics of satellite cloud picture, improve the accuracy and stability of cloud track prediction, with high scientificity and practical value.
Need to check novelty before this filing date? Find Prior Art